Anderson-Business Process Improvement Toolbox

Business Process Improvement Toolbox Second Edition

Also available from ASQ Quality Press: Root Cause Analysis: Simplified Tools and Techniques, Second Edition Bjørn Andersen and Tom Fagerhaug Bringing Business Ethics to Life: Achieving Corporate Social Responsibility Bjørn Andersen Performance Measurement Explained: Designing and Implementing Your State‑of-the-Art System Bjørn Andersen and Tom Fagerhaug Managing Service Delivery Processes: Linking Strategy to Operations Jean Harvey Integrating Inspection Management into Your Quality Improvement System William D. Mawby Enterprise Process Mapping: Integrating Systems for Compliance and Business Excellence Charles G. Cobb The Process-Focused Organization: A Transition Strategy for Success Robert A. Gardner Decision Process Quality Management William D. Mawby The Process Auditing Techniques Guide J.P. Russell Defining and Analyzing a Business Process: A Six Sigma Pocket Guide Jeffrey N. Lowenthal The Certified Manager of Quality/Organizational Excellence Handbook, Third Edition Russell T. Westcott, editor Making Change Work: Practical Tools for Overcoming Human Resistance to Change Brien Palmer To request a complimentary catalog of ASQ Quality Press publications, call 800‑248-1946, or visit our Web site at http://www.asq.org/quality-press.

Business Process Improvement Toolbox Second Edition

Bjørn Andersen

ASQ Quality Press Milwaukee, Wisconsin

American Society for Quality, Quality Press, Milwaukee 53203 © 2007 by American Society for Quality All rights reserved. Published 2007 Printed in the United States of America 12  11  10  09  08  07   5  4  3  2  1 Library of Congress Cataloging-in-Publication Data Andersen, Bjorn.   Business process improvement toolbox / Bjorn Andersen.—2nd ed.     p. cm.   ISBN 978-0-87389-719-8 (alk. paper)   1.  Reengineering (Management)  2.  Benchmarking (Management)    3.  Reengineering (Management)—Charts, diagrams, etc.  I.  Title.   HD58.87.A53 2007   658.4′063—dc22

2007015967

No part of this book may be reproduced in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher. Publisher: William A. Tony Acquisitions Editor: Matt Meinholz Project Editor: Paul O’Mara Production Administrator: Randall Benson ASQ Mission: The American Society for Quality advances individual, organizational, and community excellence worldwide through learning, quality improvement, and knowledge exchange. Attention Bookstores, Wholesalers, Schools, and Corporations: ASQ Quality Press books, videotapes, audiotapes, and software are available at quantity discounts with bulk purchases for business, educational, or instructional use. For information, please contact ASQ Quality Press at 800-248-1946, or write to ASQ Quality Press, P.O. Box 3005, Milwaukee, WI 53201-3005. To place orders or to request a free copy of the ASQ Quality Press Publications Catalog, including ASQ membership information, call 800-248-1946. Visit our Web site at www.asq.org or http://qualitypress.asq.org.

  Printed on acid-free paper

Contents List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Chapter 1­  Business Process Improvement . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Why Is Improvement Necessary? . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 A Business Process Improvement Framework . . . . . . . . . . . . . . . . .

1 1 4

Chapter 2 Understanding the Organization’s Stakeholders and Strategic Direction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Stakeholder Expectations of Improvement . . . . . . . . . . . . . . . . . . . . 2.2 Reviewing the Organization’s Strategy . . . . . . . . . . . . . . . . . . . . . . . 2.3 SWOT Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Competitive Forces Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Strategy Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9 9 17 18 20 22

Chapter 3  Understanding Your Current Business Processes . . . . . . . . . 3.1 From Processes to Departments to Business Processes . . . . . . . . . . 3.2 Definition of a Business Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Classification of Business Processes . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Identifying the Organization’s Business Processes . . . . . . . . . . . . . . 3.5 Business Process Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Relationship Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7 Traditional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8 ­ Cross-­Functional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.9 Flowchart Divided into Process Areas . . . . . . . . . . . . . . . . . . . . . . . . 3.10 ­ Several-­Leveled Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.11 Flowchart with Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.12 Paper and Pencil or PC? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27 27 32 34 37 38 43 46 51 54 55 58 62



vi  Contents

Chapter 4 Using Performance Measurement in Business Process Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 The Role of Performance Measurement in Business Process Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Implementing a Performance Measurement System . . . . . . . . . . . . . 4.3 Performance Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Hard versus Soft Indicators . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.2 Financial versus Nonfinancial Indicators . . . . . . . . . . . . . . . . 4.3.3 Leading versus Lagging Indicators . . . . . . . . . . . . . . . . . . . . .

65 65 67 70 70 71 72

Chapter 5  Creating a Business Process Improvement Road Map . . . . . 5.1 Improvement Project Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Spider Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 Performance Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Criteria Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Quality Function Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75 75 76 78 81 83 85

Chapter 6  Organizing for Business Process Improvement . . . . . . . . . . 6.1 Business Process Improvement Skills . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Organizational Modes That Support Business Process Improvement 6.3 Quality Circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Cross-­Functional Teams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Stimulating an Improvement Culture . . . . . . . . . . . . . . . . . . . . . . . .

93 93 95 100 102 103

Chapter 7  The Business Process Improvement Toolbox . . . . . . . . . . . . . 7.1 The Need for a Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 A Business Process Improvement Model . . . . . . . . . . . . . . . . . . . . . 7.3 The Tools in the Toolbox . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

107 107 108 111

Chapter 8 Tools for Collecting Data about the Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Surveying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Check Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Problem Concentration Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . .

113 113 115 117 119

Chapter 9  Tools for Analyzing the Performance Shortcoming . . . . . . . 9.1 Critical Incident . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Pareto Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Cause-­and-Effect Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Five Whys Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123 123 125 128 132

Contents  vii

9.5 Scatter Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6 Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.7 Relations Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.8 Matrix Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.9 Is–Is Not Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.10 Bottleneck Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

134 136 142 145 148 151

Chapter 10 Tools for Generating Ideas and Choosing among Them . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1 Brainstorming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Brainwriting/Crawford Slip Method . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Six Thinking Hats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4 Nominal Group Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 Paired Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

157 157 159 161 163 165

Chapter 11  Tools for Creating Improvements . . . . . . . . . . . . . . . . . . . . . 11.1 Streamlining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.1 Bureaucracy Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.2 Redundancy Elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.3 ­ Value-­Added Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.4 Process Cycle Time Reduction . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Idealizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 QFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4 Statistical Process Control/Control Chart . . . . . . . . . . . . . . . . . . . . . 11.4.1 Definitions of Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4.2 Basic Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4.3 Types of Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4.4 Constructing Control Charts . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4.5 Interpreting the Control Charts . . . . . . . . . . . . . . . . . . . . . . . . 11.4.6 Process Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5 Six Sigma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5.1 The Six Sigma Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.5.2 Six Sigma in the Organization . . . . . . . . . . . . . . . . . . . . . . . . 11.6 Business Process Reengineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6.1 A Process for Conducting BPR . . . . . . . . . . . . . . . . . . . . . . . . 11.6.2 Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6.3 Reengineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6.4 Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6.5 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.7 Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.7.1 Definitions of Benchmarking . . . . . . . . . . . . . . . . . . . . . . . . . 11.7.2 Conducting a Benchmarking Study . . . . . . . . . . . . . . . . . . . .

167 167 168 169 169 172 177 179 182 183 183 186 189 196 200 204 204 206 207 209 210 211 215 216 221 221 227

viii  Contents

Chapter 12  Tools for Implementing Improvements . . . . . . . . . . . . . . . . 12.1 A∆T Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Tree Diagram and Process Decision Program Chart . . . . . . . . . . . . . 12.3 Force Field Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

237 239 241 245

Chapter 13  Sample University—Improving Student Satisfaction . . . . . 13.1 Description of the Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Development of Performance Priorities . . . . . . . . . . . . . . . . . . . . . . 13.3 Understanding the Current Processes and Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.4 Analysis of the Performance Shortcoming . . . . . . . . . . . . . . . . . . . . 13.5 Generating Ideas and Improvement Proposals . . . . . . . . . . . . . . . . . 13.6 Implementation of Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . .

251 251 252

Chapter 14  Green Carpet Seed—Revamping the Business Model . . . . 14.1 Description of the Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Development of Performance Priorities . . . . . . . . . . . . . . . . . . . . . . 14.3 Understanding the Current Processes and Performance Shortcoming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.4 Generating Improvement Ideas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.5 Developing an Improved Process . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.6 Implementation of the New Service . . . . . . . . . . . . . . . . . . . . . . . . .

263 263 264

Appendix  Template Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SWOT Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spider Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Performance Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Criteria Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality Function Deployment (QFD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relationship Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-Functional Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Check Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pareto Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cause-and-Effect Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Five Whys Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scatter Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relations Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Is–Is Not Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paired Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Control Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Force Field Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

275 276 276 277 277 278 279 280 280 281 281 282 282 283 283 284 284 285 286 286

254 257 258 260

266 268 269 271

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

List of Figures Figure 1.1

Without maintenance and improvement, the performance level decreases. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 1.2 Business process improvement framework. . . . . . . . . . . . . . . Figure 1.3 The Deming wheel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.1 The stakeholder model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.2 Stakeholder classification matrix. . . . . . . . . . . . . . . . . . . . . . . Figure 2.3 The Kano model and the three types of stakeholder requirements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.4 Stakeholder strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.5 SWOT analysis results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.6 Five competitive forces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.7 Generic strategy map. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2.8 Strategy map for the shipping company. . . . . . . . . . . . . . . . . . Figure 3.1 The contradiction between vertical departments and horizontal processes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.2 A simple business process. . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.3 The same simple business process, now with department boundaries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.4 Primary and support processes. . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.5 Business processes in ENAPS. . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.6 Business processes derived from strategy and stakeholders. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3.7 The hierarchy of process modeling.. . . . . . . . . . . . . . . . . . . . . Figure 3.8 Example of a relationship map. . . . . . . . . . . . . . . . . . . . . . . . Figure 3.9 The resulting relationship map for the supply process. . . . . . Figure 3.10 Flowchart for the supply process. . . . . . . . . . . . . . . . . . . . . . . Figure 3.11 Flowchart for the filing process. . . . . . . . . . . . . . . . . . . . . . . . Figure 3.12 Standard IDEF0 diagram design. . . . . . . . . . . . . . . . . . . . . . .

ix

3 5 7 12 12 13 15 19 20 23 24 28 29 30 35 36 37 41 44 46 47 49 50

  List of Figures

Figure 3.13 Figure 3.14 Figure 3.15 Figure 3.16 Figure 3.17 Figure 3.18 Figure 3.19 Figure 3.20 Figure 4.1 Figure 4.2 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 Figure 5.8 Figure 5.9 Figure 5.10 Figure 5.11 Figure 5.12 Figure 5.13 Figure 5.14 Figure 5.15 Figure 6.1 Figure 6.2 Figure 6.3 Figure 7.1 Figure 8.1 Figure 8.2 Figure 8.3 Figure 9.1 Figure 9.2 Figure 9.3 Figure 9.4 Figure 9.5 Figure 9.6 Figure 9.7 Figure 9.8

 ample IDEF0 diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S Example of a ­cross-­functional flowchart. . . . . . . . . . . . . . . . . Cross-functional flowchart for the reporting process. . . . . . . Example of a flowchart divided into process areas. . . . . . . . . Level 0 chart for the process. . . . . . . . . . . . . . . . . . . . . . . . . . Level 1 chart for procurement. . . . . . . . . . . . . . . . . . . . . . . . . Level 1 chart for reporting. . . . . . . . . . . . . . . . . . . . . . . . . . . . Flowchart with bank customer movement statistics. . . . . . . . The performance measurement system design process. . . . . . Performance measurement as an early warning system. . . . . Trend analysis for three example measures. . . . . . . . . . . . . . . Results of the trend analysis. . . . . . . . . . . . . . . . . . . . . . . . . . Example of a spider chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . Spider chart for the central performance measures. . . . . . . . . A generic performance matrix. . . . . . . . . . . . . . . . . . . . . . . . . Performance matrix for the six performance measures. . . . . . Matrix for criteria testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Criteria testing for a hair salon chain. . . . . . . . . . . . . . . . . . . . The basic structure of QFD. . . . . . . . . . . . . . . . . . . . . . . . . . . A chain of QFD charts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The house of quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benchmarking of others’ offers. . . . . . . . . . . . . . . . . . . . . . . . Symbols for the relational matrix. . . . . . . . . . . . . . . . . . . . . . Symbols for the roof matrix. . . . . . . . . . . . . . . . . . . . . . . . . . . QFD chart for an internal IT department. . . . . . . . . . . . . . . . . Traditional organization with vertical silos. . . . . . . . . . . . . . . Organizing along business processes with process owners. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Organization of quality circles. . . . . . . . . . . . . . . . . . . . . . . . . Stages in business process improvement work. . . . . . . . . . . . Example of a check sheet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Check sheet for recording the causes of lost bids. . . . . . . . . . Cutout of city map with streetlight problem concentration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Critical incidents for the switchboard operators. . . . . . . . . . . A general Pareto chart with a line for cumulative importance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pareto chart for value of lost jobs. . . . . . . . . . . . . . . . . . . . . . The structure for a fishbone chart. . . . . . . . . . . . . . . . . . . . . . Process chart with fishbone charts for each process step. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fishbone chart for shaft manufacturing. . . . . . . . . . . . . . . . . . List representation of five whys. . . . . . . . . . . . . . . . . . . . . . . . Five whys analysis chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50 52 53 55 56 57 58 61 69 73 77 78 79 80 82 83 84 85 86 87 87 88 88 88 91 96 97 102 109 118 119 121 125 126 127 128 130 131 132 133

List of Figures  xi

Figure 9.9 Figure 9.10 Figure 9.11 Figure 9.12 Figure 9.13 Figure 9.14 Figure 9.15 Figure 9.16 Figure 9.17 Figure 9.18 Figure 9.19 Figure 9.20 Figure 9.21 Figure 9.22 Figure 9.23 Figure 9.24 Figure 9.25 Figure 9.26 Figure 9.27 Figure 9.28 Figure 10.1 Figure 10.2 Figure 10.3 Figure 11.1 Figure 11.2 Figure 11.3 Figure 11.4 Figure 11.5 Figure 11.6 Figure 11.7 Figure 11.8 Figure 11.9 Figure 11.10 Figure 11.11 Figure 11.12 Figure 11.13 Figure 11.14 Figure 11.15 Figure 11.16 Figure 11.17

 catter charts showing different degrees of correlation. . . . . . S Scatter chart to analyze weather and jobs. . . . . . . . . . . . . . . . Check sheet for the example. . . . . . . . . . . . . . . . . . . . . . . . . . Histogram for the example. . . . . . . . . . . . . . . . . . . . . . . . . . . . Centered process with a small variation width. . . . . . . . . . . . Centered process with a large variation width. . . . . . . . . . . . . Process with a small variation width, but not centered. . . . . . Histogram with two peaks. . . . . . . . . . . . . . . . . . . . . . . . . . . . Histogram that is cut off. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The principles of a qualitative relations diagram. . . . . . . . . . . Qualitative relations diagram for a poorly functioning measurement system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A general quantitative relations diagram. . . . . . . . . . . . . . . . . Quantitative relations diagram for the example. . . . . . . . . . . . Types of matrix diagrams. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relations symbols. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Matrix diagram for the video rental chain. . . . . . . . . . . . . . . . Generic table for is–is not analysis. . . . . . . . . . . . . . . . . . . . . Ski resort is–is not analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . A bottleneck resource limits the flow from the entire process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emergency ward bottleneck analysis. . . . . . . . . . . . . . . . . . . . Individual ranking card. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Resulting flip chart list. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paired comparisons matrix for internal communication. . . . . ­Value-­added analysis questions. . . . . . . . . . . . . . . . . . . . . . . . ­Cost-­cycle time chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Corresponding chart after completed ­value-­added analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A flowchart with steps marked according to streamlining results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ­Cost-­cycle time chart for the original process. . . . . . . . . . . . . ­Cost-­cycle time chart for the streamlined process. . . . . . . . . . Analysis of the current and ideal process in idealizing . . . . . . Idealizing in a printing company. . . . . . . . . . . . . . . . . . . . . . . The general house of quality, the chart used by QFD. . . . . . . House of quality for process improvement. . . . . . . . . . . . . . . Normal distribution curve. . . . . . . . . . . . . . . . . . . . . . . . . . . . Types of control charts and when to use them. . . . . . . . . . . . . – Example of an X chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Example of an R chart. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Control chart with one point outside the control limits. . . . . Chart with several points outside the control limits. . . . . . . . . Process with change in level. . . . . . . . . . . . . . . . . . . . . . . . . .

134 136 138 138 140 140 141 141 142 142 143 144 145 146 147 148 149 150 152 154 165 165 166 171 172 173 174 176 176 177 178 180 181 185 188 191 192 197 198 198

xii  List of Figures

Figure 11.18 A  series of 12 or 14 points on the same side of the center line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.19 Process with trend. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.20 Process probably involving tampering. . . . . . . . . . . . . . . . . . . Figure 11.21 Chart displaying cycles. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.22 p chart for one day’s production. . . . . . . . . . . . . . . . . . . . . . . Figure 11.23 Control chart divided into operators. . . . . . . . . . . . . . . . . . . . Figure 11.24 The process for conducting a BPR project. . . . . . . . . . . . . . . Figure 11.25 Process design using the clean sheet approach. . . . . . . . . . . . Figure 11.26 The current process for developing a tailor-­made system. . . Figure 11.27 The ideal process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.28 Operational definition of benchmarking. . . . . . . . . . . . . . . . . Figure 11.29 Benchmarking out of the box. . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.30 Three types of benchmarking based on what is being compared. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.31 Recommended combinations of types of benchmarking. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.32 Progress of a typical benchmarking study. . . . . . . . . . . . . . . . Figure 11.33 Models for organizing a benchmarking study. . . . . . . . . . . . . Figure 11.34 The benchmarking process based on the benchmarking wheel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11.35 Ambition pyramid for benchmarking partners. . . . . . . . . . . . Figure 11.36 Three levels of benchmarking information. . . . . . . . . . . . . . . Figure 11.37 Methods and tools for the observation phase. . . . . . . . . . . . . . Figure 12.1 Matrix for assessing implementation order. . . . . . . . . . . . . . . Figure 12.2 Flowchart for the A∆T analysis. . . . . . . . . . . . . . . . . . . . . . . . Figure 12.3 A principal tree diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 12.4 Tree diagram for introducing a computer system. . . . . . . . . . Figure 12.5 A general PDPC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 12.6 PDPC for the library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 12.7 Force field diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 12.8 Force field analysis for a new library computer system . . . . . Figure 13.1 Spider chart for external comparison. . . . . . . . . . . . . . . . . . . . Figure 13.2 Performance matrix for establishing priorities. . . . . . . . . . . . Figure 13.3 Flowchart for the course process. . . . . . . . . . . . . . . . . . . . . . . Figure 13.4 Fishbone chart for lacking communication ability. . . . . . . . . . Figure 13.5 Matrix diagram for Sample University. . . . . . . . . . . . . . . . . . Figure 13.6 Force field analysis for the implementation. . . . . . . . . . . . . . . Figure 14.1 Criteria testing matrix for Green Carpet Seed. . . . . . . . . . . . . Figure 14.2 ­Cross-­functional flowchart for the dealer management and support process. . . . . . . . . . . . . . . . . Figure 14.3 PDPC for Green Carpet Seed. . . . . . . . . . . . . . . . . . . . . . . . . .

199 199 200 200 203 203 210 215 219 220 222 224 225 225 226 226 228 230 231 232 238 240 242 243 244 245 247 248 253 254 255 257 258 261 266 267 273

List of Tables Table 4.1 Table 4.2 Table 7.1 Table 9.1 Table 9.2 Table 9.3 Table 11.1 Table 11.2 Table 11.3 Table 11.4

Differences between hard and soft indicators. . . . . . . . . . . . . Examples of financial indicators. . . . . . . . . . . . . . . . . . . . . . . Tools for different phases in the improvement work. . . . . . . . Measures of the diameter of a hole for 125 work pieces. . . . . Determining the number of classes, C. . . . . . . . . . . . . . . . . . . Formats with the number of variables and relations for matrix diagrams. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Calculation factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Factors for estimating the standard deviation. . . . . . . . . . . . . Focus areas for systematic reengineering. . . . . . . . . . . . . . . . Breakthroughs through benchmarking. . . . . . . . . . . . . . . . . . .

xiii

70 72 110 137 137 147 190 201 212 224

Preface

I

n the years since the first edition of this book was published, in 1999, I have received a great deal of feedback from readers. The book has been used as curriculum in many different courses, including some that I teach. The stu‑ dents have responded positively, and those without much work experience find the numerous examples useful. The larger bulk of readers are, I believe, people who work with process improvements in their jobs. They seem to come from all types of sectors and industries, and again the feedback has been predominantly positive. I never intended this book to be an academic text that would further the state of the art in this field, and this is probably one reason for the book’s reception. There are between 12,000 and 15,000 ­first-­edition copies in circulation. From time to time, I flip through the book, either to review some topic before a lecture or to explain something to students in more detail. Whenever I do, I cannot help noticing many shortcomings in the first edition. I have learned a lot about improvement work during these years, and the general knowledge about these issues has increased significantly. In the end, it became clear that a revision was in order if the book is to remain useful. The second edition contains quite a few extensive changes. First, the entire structure of the first half of the book has been reworked. I was never too happy with the overall improvement model that formed the basis for the first edition, and I believe this is much clearer now. Material on how to organize improvement work and create a culture that supports a relentless quest for improvement has been added, as have additional details on strategic and performance planning. Some new tools and techniques have been added to the toolbox, and a new, extensive case study, Green Carpet Seed, should help facilitate the more coherent use of the toolbox. As for the first edition, the objective is still to give readers practical insight into how they can create a coherent business process improvement system. I genuinely believe that consistently working on improving various aspects of how things are done, large and small, is the key to success for any organization. If you can com‑

xv

xvi  Preface

bine a culture that never thinks the organization is where it needs to be and a ­well­stocked toolbox, you will have a good chance of a successful future. The first half of the book presents an overall business process improvement model. The ensuing chapters deal with understanding and modeling your current business processes, using performance measurement in improvement work, creat‑ ing a business process improvement road map, and organizing for improvement work. The second half of the book presents the overall toolbox, and each phase of the overall improvement model is discussed in its own chapter. For each of these phases, a selection of suitable tools is presented, which includes information on each tool’s background, steps for using the tool, and an example of its use. The last two chapters contain two more extensive case studies that illustrate use of the full methodology. And finally, a number of templates that support most of the tools presented can be found in the appendix. I should point out that the book is suited for employees and managers at any organizational level in any type of industry, including service, manufacturing, and the public sector. It is also useful as a textbook for students in courses relating to quality management and continuous improvement. My sincere thanks go to all those who have inspired the writing of this book, including colleagues, classroom participants in courses, and companies with which I have worked. Thanks also to my wife, Hilde Mentzoni Andersen, for reviewing the draft. And finally, I must express my gratitude to all the great people at ASQ Quality Press, who have performed excellently at every stage of the process.

Bjørn Andersen Trondheim, Norway, March 1, 2007

1 Business Process Improvement

T

o some, the term improvement is well understood and almost second nature. I guess I fall into that category, but I also understand that continually think‑ ing about improvement is more alien to others. It therefore seems logical to begin with an introduction to improvement in general.

1.1  Why Is Improvement Necessary? Take almost any area of life or any industrial sector. If you trace its history, you’ll realize that dramatic changes have occurred, most likely in a relatively short time span. These changes apply to the products or services offered in the marketplace, the technologies used, the way activities are organized, and so on. For example, consider travel. Decades ago, only a very small portion of the population, regardless of which part of the world you look at, had access to or the means to travel. Air travel was prohibitively expensive for most people, and rail and sea travel were too slow to allow people to go great distances. The only mode of transportation available to large parts of the population in some countries was the automobile. Although there is room for improvement in the world of travel, there is no doubt that there have been significant changes. In most parts of the world, air travel is now quite affordable. People are traveling everywhere, and more people are seeing distant shores than at any other time in history. From an environmental point of view, travel has probably become too accessible. Or take telecommunication. I remember when my parents decided it was too cumbersome to walk two blocks to the nearest phone booth to make a call, so they had a phone line from the ­state-­owned monopoly provider installed in the house. This was in the 1970s. The waiting time was about four months, and the phone you got came in ­standard-­issue gray, with no other options available. Today, no one would accept this type of customer service, and fortunately we don’t have to.



  Chapter One

In most countries, telecommunication, along with many other basic services that used to be run by the state, has been deregulated. As customers, we can choose our supplier, and the range of services has grown immensely. I have traveled a bit over the last few years, and I am still amazed by how easy it is to turn on my cell phone or computer with a wireless Internet card and have both work. My last trip took me to Venezuela, and I had been told that in order to use my cell phone, I’d have to buy a separate SIM card from a local provider. But when I arrived, both the phone and the ­high-­speed Internet worked right away. Turns out the information I had was two months old. There have also been considerable changes in the field of electronics. Do you own a DVD player, or maybe a DVD recorder, with a hard drive of x gigabytes? Perhaps you own an MP3 player? Many of us own one or more of these units, and we take it for granted that new technology will emerge before we wear out our cur‑ rent equipment. When I bought my first CD burner, I knew it was only a matter of time before the DVD burner was available. And when I got my first DVD burner, the consumer electronics industry had already made it clear that ­dual-­layer burners were imminent. Now, the first ­Blu-­ray Disc players are available, and I will bet you anything they are not the last evolutionary step in optical storage technology. What is remarkable is the price development for these types of products. The first DVD players cost from $2000 to $3000 (luckily, having understood how these things work, I avoided buying at that time). Today, you can pick up a DVD player right next to light bulbs in any chain store for $49.99. These are only a few examples, but I daresay you’ll find the same trend no mat‑ ter where you look. In fact, I’d say the adage of “the good old days” is simply not true. Granted, I’m old enough to claim this myself now and then, usually referring to ­hip-­hop music or the lack of exercise by kids these days, but for the most part, things are actually much better today. We say that people were much more polite “in the old days,” but I generally find customer service to have improved in almost any industry I can think of. We also like to say that people used to lead much better lives, with less stress and much more time to enjoy themselves, but this is not true. Most people spent much more time earning enough money to just get by, or they spent their time doing household chores that appliances do for us today. At no point in history have people had wider choices of what to do with their lives and better chances of fulfilling their dreams and needs than they do today. Why? Because there is an ingrained human quest for what’s better, what’s around the next bend. It is this drive that has led individuals to explore new con‑ tinents, to travel into space, and to go below the surface of the oceans. This same drive has led to countless inventions, including four generations of DVD burners. If you agree with this argumentation, there simply is no way of curbing improve‑ ment efforts in organizations; people will keep going forward no matter what. In addition, many forces, both internal and external, have caused improve‑ ments to become a necessity in today’s marketplace:

Business Process Improvement  

• The performance level of most processes shows a tendency to decrease over time unless forces are exerted to maintain it. This means that to maintain the current standards, it is necessary to perform some degree of maintenance. If we want to create improvement and renewal, this requires efforts beyond pure maintenance, as shown in Figure 1.1. • If an organization does not improve, you can be quite certain its competi‑ tors will. Should the unlikely scenario occur that neither an organization nor its competitors improve, there are always other actors willing to enter the market segment. • Today’s customers are more demanding and, frankly, are spoiled. Sup‑ ply and the quality of the supply are ever increasing, which in turn causes expectations to rise dramatically. If it is not possible to exceed expectations, which is the ideal situation, they at least have to be met. If they are not met, you are guaranteed to lose the customer. Generally, this means that what was satisfactory a few years ago barely passes today, and quite certainly will soon be below expectations. It is therefore irrelevant to discuss whether we have to improve; the question, rather, is how much do we improve and how quickly do we do it.

Performance level

Breakthrough

Improvement Maintenance No improvement effort Time Figure 1.1  Without maintenance and improvement, the performance level decreases.

  Chapter One

There is, of course, no definitive answer to this, but a rather general one is that continuous improvement, combined with occasional breakthroughs, is needed. In this respect, it is worth pointing out that experience indicates that an organization that emphasizes continuous improvement usually also pos‑ sesses the creativity and attitude toward improvement necessary to create an occasional breakthrough (Hayes and Wheelwright, 1984). The opposite, trying exclusively to achieve a breakthrough and neglecting the continuous improve‑ ment activity, might prove difficult in creating a breakthrough when one is needed. After observing a large number of improvement projects in very different types of organizations, I have found that people differ greatly when it comes to innate improvement skills. Some people, albeit only a few, are born pioneers or inventors or architects of ­large-­scale change, and some organizations seem to be natural champions at continuous improvement. Sadly, although all of us share humankind’s drive toward progress, many of us have discovered that improvement is a skill that must be learned and practiced. I have also realized how much easier it is to exert this skill when it’s linked to something specific, such as a ­well-­defined improvement project or the use of an improvement tool, where certain steps that need to be executed are outlined. Regardless of how you organize the improvement efforts and which tools you use, business processes are a good starting point for improvement work. Together, these observations converge into the main idea of this book, giving readers a ­well-­stocked toolbox for use in improving their organi‑ zations’ business processes.

1.2  A Business Process Improvement Framework Although I argue that a drive toward improvement is ingrained in humankind, I am not saying that improvement work that is goal oriented, includes people, and performs well in an organization is easy. There are a number of prerequisites that need to be in place to succeed in this game. While the diagram in Figure 1.2 might not be 100 percent complete, I have tried to portray what I believe are the most important elements an organization needs to understand or master to perform well with its improvement efforts. In the spirit of the topic of the book, this is an updated, hopefully improved, version of an improvement model pre‑ sented in the first edition. The old model outlined a sequence of activities that naturally belonged to a comprehensive improvement process. Some of these elements also prevail in the new version, but in the years since the first edition, I’ve come to realize that there are a number of other elements that warrant men‑ tion in such a framework. The elements of the business process improvement framework are the following:

Business Process Improvement  

• The organization’s stakeholders and its strategy, which are included because they form the overall priorities of the organization. They define the critical areas where the organization needs to excel, and, even if this is not always the case, they should be the backdrop against which any improvement proj‑ ect ideas are assessed. • A thorough understanding of the current business processes. I have already argued that business processes are an ideal unit of analysis for improvement work, and understanding how a process is currently performed is an invalu‑ able platform for improving it. • Performance measurement, which plays an integral part in improvement work. It helps assess which business processes should be improved, it gives insight into the current execution of the process, and it allows tracking of the improvement effects created. • A business process improvement road map, which is a type of game plan for the improvement efforts. Such a road map will typically outline which busi‑ ness processes need to be improved over the long term, and it will contain details about more immediate projects. • An organizational structure with skills, incentives, and attitudes that foment improvement. This is in itself a vast topic and an area of critical importance for an organization that has high improvement ambitions. If the ­employees

Stakeholders

Improvement toolbox

Strategy

Performance measurement

Improvement road map

Organization and skills for improvement Business process understanding

Figure 1.2  Business process improvement framework.

  Chapter One

don’t have basic values that support continuous improvement, or there are no incentives to invest extra effort in improvements, or no one has the responsibility of initiating improvements, the effects will be dismal. • The business process improvement toolbox, which this book is ultimately about. Time and time again, I have seen that effective, ­easy-­to-use improve‑ ment tools help set a direction, guide improvement projects on their way, and overcome a sense of not knowing how to get started with improvement work. The framework combines elements that determine the direction of the improve‑ ment efforts, provide a platform of current business understanding, and give tools and techniques for running improvement projects. The next chapters of the book deal with the elements in the framework, and the toolbox itself is presented over several chapters. What this version of the improvement framework fails to illustrate is the cycli‑ cal nature of improvement work. Improvement projects often progress through phases of assessing the current situation, creating and planning improvements, implementing the changes, and measuring whether the improvement targets were reached. If the targets were not reached, the process is often repeated until they have been achieved. This cycle applies to both single improvement projects and the organization as a whole, and the unique strength that comes with such a virtuous circle is this stubborn attitude of never being content. When this culture settles and prevails throughout the organization, there are no limits to how far it can go. Those familiar with the work of W. Edwards Deming will see that such a cyclic idea is based on the principles of the Deming wheel (Deming, 1986). The Deming wheel—or the cycle of plan, do, check, act—is shown in Figure 1.3. This control loop illustrates a general approach for conducting continuous improvement. The activities of the four phases are as follows: • Plan—In this phase, the problem is analyzed and activities to remedy it are planned, that is, improvement planning. • Do—In this phase, the activities planned in the previous phase are carried out. The purpose is primarily to experiment with the solution. • Check—In this phase, measurements are made to evaluate whether the activities had any effect on the problem, that is, a performance evaluation. • Act—Finally, in this phase, the process is modified according to the activities that were confirmed to give results. After performing the four phases, you will have either an improved process or a starting point for a new cycle.

Business Process Improvement  

Act

Plan

Check

Do

Figure 1.3  The Deming wheel.

The purpose of the Deming wheel, besides describing a systematic approach to improvement, is that the wheel should continuously be in motion by performing this process. This is one of the main messages of this book and one that will be repeated several times. The next chapter deals with the organization’s stakeholders and its strategy, which is one area where this relentless pursuit of improvement must be emphasized.

2 Understanding the Organization’s Stakeholders and Strategic Direction

Y

ou have not misread the title of this book; it is about business process improvement. Although the book is not about strategic planning, the strat‑ egy of the organization strongly impacts its improvement efforts. This chapter explains this influence and presents a couple of useful techniques at this preimprovement stage.

2.1  Stakeholder Expectations of Improvement There are many ways of presenting improvement ideas and proposals for improve‑ ment projects in an organization. In the 1970s and perhaps partway into the 1980s, quality circles were a popular approach for bringing people together to discuss and develop improvement ideas. In many organizations, mailboxes (physical or elec‑ tronic) have been used to collect ideas and suggestions from employees. Likewise, feedback from customers can be collected more or less systematically through forms in stores or offices, through an e-mail service, or through a discussion Web page. Also, managers for each area or level of the organization can be responsible for recommending a number of improvement ideas per year. In most cases, several mechanisms will be in play. The message I want to convey is this: While a richness of improvement ideas and enthusiasm in the organization for creating ideas is extremely important, a bal‑ ance must be struck where ideas sanctioned for further investigation or implemen‑ tation are coordinated. An organization, no matter its size, will have only so many free resources—that is, time, people, and money—that can be invested in carrying out improvement projects. Keeping the operations going will normally take prior‑ ity, leaving improvement work as an extra effort that is often sacrificed when other tasks require attention. To ensure that the return on the improvement investments



10  Chapter Two

is maximized, projects should address areas of the business that will make a differ‑ ence to the stakeholders. Following this logic, an important platform for creating the improvement road map, mentioned in the previous chapter, is a solid understanding of who the orga‑ nization’s stakeholders are, what expectations they hold, what they are satisfied with, and in which areas they would like to see the organization improve. In other words, a stakeholder analysis, a process that now seems to be fairly well known and widespread in most industrial sectors, should be carried out. Stakeholder analysis can easily be defined as a tool in strategic planning, but for improvement purposes, it also serves other useful functions: • When conducted at a higher level, that is, for the entire organization or busi‑ ness unit, it creates an overall understanding of the organization’s position on the competitive map of the sector it operates in. This allows top man‑ agement to better understand the direction it should follow to maintain or strengthen this position, thereby providing a basis for prioritizing improve‑ ment projects. • Performing a stakeholder analysis once or twice per year can in itself gen‑ erate improvement ideas. Depending on the level of detail of the analysis, discussions about, or even with, key stakeholders can take place regarding their satisfaction with the organization and what they would like to see done differently. • Even in individual improvement projects, a ­lower-­level stakeholder analy‑ sis can be useful, especially in the beginning of a project. Consider a small group of health club employees who have been given the task of reversing the decreasing turnover at the ­in-­house juice bar over the last few months. A limited stakeholder analysis for the juice bar would quickly give the group useful insight into members’ refreshment preferences, how members spend their time in the club, and the incentives that the employees have for offer‑ ing good service. It would most likely have to be followed up by other analysis tools, but it would still be a good starting point. You may have noticed that I use the word stakeholder with the expectation that readers already know what it means. I hope this is justified, and I know that many fellow authors take the same position, as the term has become highly entrenched in management literature. Over the years, I have, however, tried to figure out where the term really comes from, and I am not sure I have succeeded yet. Wikipedia (2006) is usually reliable, and the explanation found there links the word to a tradi‑ tional usage from law and gambling: “a third party who temporarily holds money or property while its owner is still being determined.” This might very well be true, and it might precede a theory I heard recently, related to the American pioneers

Understanding the Organization’s Stakeholders and Strategic Direction  11

exploring the western U.S. territories during the 1800s. Having enjoyed movies and TV series like Far and Away, Centennial, and How the West Was Won over the years, and lately Into the West, I find this theory fascinating, perhaps even credible: The pioneers went west to find a new life, most of them seeking land where they could establish homes and farms. Land was claimed by planting a flag to mark a plot until a formal deed was issued. The flag was attached to a stake. Thus, accord‑ ing to this view, a stakeholder was a person setting out to claim land. I should also mention that stakeholder analysis is not one unified, ­generally ­agreed-upon tool with certain steps to be performed. It is more of a commonsense exercise aimed at generating a higher awareness of the various stakeholders who have a vested interest in the organization and, at a minimum, the needs, require‑ ments, and expectations these stakeholders have. Over the years, I have extensively studied and worked with stakeholder analysis in many enterprises, and the fol‑ lowing seems to me like a very useful and workable approach (more details about these steps follow after the list):

1. Identify the organization’s stakeholders, normally through a brainstorming effort.



2. Group or classify the stakeholders according to some criteria to determine which are the most important.



3. Identify the needs and expectations held by the stakeholders, an exercise for which the Kano model has proved suitable.



4. Try to anticipate the likely behavior of the stakeholders toward the organization.



5. With the insight generated through these steps, develop strategies for how to handle the stakeholders.

The only feasible way I have encountered for identifying stakeholders is to conduct a good ­old-­fashioned brainstorming (see section 10.1 for more details about brain‑ storming). If a small group of people from different areas of the organization gets together, it normally comes up with the main stakeholders in an hour or two. Who‑ ever undertakes this task of identifying the stakeholders should remember that there are some obvious and not so obvious stakeholders out there. The obvious ones are the customers, suppliers, competitors, employees, and perhaps the owners. Some that are easy to overlook include the partners, the media, government/authorities at various levels and of different types, the local community, and a number of pressure groups. Figure 2.1 illustrates how some of these stakeholders surround the organi‑ zation and constitute important parts of the organization’s environment. You can find several models for classifying stakeholders; see, for example, D’Herbemont and Cesar (1998). One that I find useful assesses each stakeholder’s

12  Chapter Two

Authorities

Owners

Customers

Suppliers Financial institutions

Employees

Management

Competition

Environment

Alliance partners

Figure 2.1  The stakeholder model.

potential for impacting and cooperating with the organization, as shown in Fig‑ ure 2.2 (Savage, Nix, Whitehead, and Blair, 1991). The total set of stakeholders will be scattered across this matrix. In general, the farther to the left and top of the matrix, that is, toward “high” on both dimensions, the more important the stakeholders typically will be for the organization. All these stakeholders hold certain expectations toward the organization. Iden‑ tifying these expectations and translating them into more tangible quantities can probably be done by following different avenues. The organization most likely already possesses a fairly high level of knowledge about some stakeholders’ requirements, for example, owners or customers. For others, it might be difficult to tell what they do or do not want. For pressure groups advocating equal opportuni‑ ties, the media in general, or public authorities, the expectations are probably much less clearly understood, and hidden agendas may even exist. If you map all of the probable and improbable expectations your stakeholders harbor, you will undoubt‑ edly discover that these are plentiful, not always coherent, and very wide ranging. To bring some order to these expectations and be able to differentiate between the important and the not-so-important ones, a very useful diagram called the Kano model is at your fingertips (Kano, Seraku, Takahashi, and Tsuji, 1984). As you can

High

High

Low

Potential for cooperation with the organization

Potential to impact the organization

Figure 2.2  Stakeholder classification matrix.

Low

Understanding the Organization’s Stakeholders and Strategic Direction  13

see in Figure 2.3, the model is nothing more than an ­awareness-­creating diagram showing that there are different types and levels of stakeholder requirements. The straight diagonal line in the figure portrays the clearly expressed require‑ ments of the stakeholder. Generally, these are the only demands the stakeholder will describe if asked about her or his wishes. If the stakeholder is a major share‑ holder of the organization, he or she could express the requirements that the return on investment be a minimum of 7.2 percent, that he or she be granted a seat on the board, and so on. In addition, there also exists a set of requirements that are so basic that they are not even expressed, as indicated by the lower curve. For the shareholder, these could be that the organization does not go broke and lose its capital, that it does not get involved in criminal or otherwise unethical activities that could harm the shareholders, that business is conducted according to general rules and customs, and so on. Together, these two sets of requirements constitute a complete set of demands imposed by the stakeholder toward the organization. The satisfaction depends on how well both sets of requirements are met. It will be of no help if the investment returns 10 percent but the shareholder is arrested for criminal activities in which Loves the organization’s output Performance quality: Expressed Measurable Excitement quality: Not known Not expressed Trend sensitive Stakeholder specific

Degree of achievement

Basic quality: Assumed Obvious Often forgotten

Hates the organization’s output Figure 2.3  The Kano model and the three types of stakeholder requirements.

14  Chapter Two

he has involved the company. In other words, satisfying expressed requirements cannot rectify shortcomings in the basic demands. On the other hand, satisfying every single one of the basic requirements will not lead to complete satisfaction unless the expressed requirements have also been fulfilled. At best, this will elimi‑ nate dissatisfaction. The danger is that the stakeholder takes it for granted that the organization is aware of the basic requirements, but this might not be the case. Such silent assumptions are one of the main focuses when clarifying requirements in the stakeholder analysis. If these two sets of requirements are defined and satisfied, the foundation for satisfaction should be firmly established. To further enhance satisfaction, and even delight the stakeholder, we can look to a third set of requirements. Requirements is not really the correct word, as these are conditions not expressed by the stakeholder, and often the stakeholder itself is not even aware of these needs. For the shareholder, this could include having the organization make all arrangements for transportation to general assemblies, free access to the organization’s products or services, a special Web site for the shareholders that presents updated information of interest to them, and so on. If both the basic and expressed requirements have been satisfied, the fulfillment of such extra “requirements” can create true delight. These are often the little extras required to ensure loyalty and access to the best stakeholders of all kinds. Note, though, that if such extra requirements are delivered once or a few times, they are often added to the expressed or even basic requirements that must be ful‑ filled to avoid dissatisfaction.

Example The financial department of a large company produced weekly reports for the sales department that contained the credit status of the custom-­ ers. As he always did, the responsible accountant put his heart into the accuracy of the report, even if this often took quite some time and could sometimes delay the report. The sales department was not depending on minute accuracy, but was depending on having the report on time to be able to issue order confirmations at the agreed time. It assumed that the financial department knew about this and did nothing to attempt to com-­ municate it to the accountants. Instead, the sales department continued to be annoyed by the delayed reports, without appreciating the accuracy, and nothing improved.

When trying to anticipate how the stakeholders will act toward the orga‑ nization, it can be useful to take the stakeholder classification matrix one step further. Figure 2.4 shows the same matrix as in Figure 2.2 but now with labels for each field.

Understanding the Organization’s Stakeholders and Strategic Direction  15

High

Low

High

Mixed blessing

Supportive

Low

Potential for cooperation with the organization

Potential to impact the organization

Nonsupportive

Marginal

Figure 2.4  Stakeholder strategies.

These labels do not represent exact predictions of behavior for each group, but they indicate from experience how each group might act. For each group, there are also some recommendations about suitable strategies to follow: • Supportive: should be involved in relevant discussions and decisions • Marginal: should simply be monitored • Nonsupportive: use a defensive strategy and minimize the dependency on the stakeholder • Mixed blessing: best handled through cooperation To recap, the outcome of the stakeholder analysis should constitute a very good start for the ensuing improvement efforts.

Example A large bank that was a leader in developing Internet banking services had pushed this approach to a point where customers had virtually no reason to see a bank employee; everything could be done electronically or, in the worst case, by phone. Two years after the project was deemed complete, figures showed that the bank was losing certain customers, especially elderly customers and customers with ­ above-­average income or investment funds. Since these were also the most profitable customers, this posed a serious challenge to the bank, and it decided to undertake a stakeholder analysis. A large number of stakeholders were identified, including authorities, investment funds, financial consultants, employees, trade unions, and so on. These stakeholders were subsequently analyzed according to ­importance

16  Chapter Two

for the bank, and since the problem addressed was loss of customers, only those found relevant in this matter were included in the process: customers in different segments (25 years and below, 25–50 years, 50–65 years, 65 years and above, and investors), financial consultants (who often advised customers about which bank to choose), and trade unions (who brokered banking deals on behalf of their members). For each of these, the Kano principles were applied to identify needs and expectations. To simplify, only selected expectations identified in the analysis of the customers of age 65 and above are included here.

Performance Expectations • Competitive interest rates • Competitive fees and expenses • Easy-to-understand information from the bank

Basic Expectations • Easy access to a customer representative when required • Active counseling regarding major financial decisions • The bank has my interests in mind

Excitement Expectations • Branch office open 24/7 • Invitations to annual events sponsored/hosted by the bank • Birthday attention • Internet banking available From this analysis, the team was surprised to find that this customer seg-­ ment took it for granted that there would be a physical branch office they could go to and where they would find people who could give them advice. Internet banking was simply a possible extra service for some of these customers. For the customers with more funds to invest, the picture was almost the same: they relied on active and close contact with a representa-­ tive from the bank to aid their financial decisions. After seriously considering continuing the policy to shift customers to electronic channels and accepting the loss of those who felt uncomfortable with this, the bank decided to change its strategy. A new brand was estab-­ lished to handle customers with greater need for personal contact. This group started with five pilot offices in which the branch office was rebuilt and tailored to the needs of these customers. The approach was a huge success, with these offices winning back many more customers than had been lost the last few years, obviously due to filling a gap that few banks

Understanding the Organization’s Stakeholders and Strategic Direction  17

concentrated on. Today, this bank has by far the highest market share in this customer segment and operates close to 50 such branch offices.

2.2  Reviewing the Organization’s Strategy I have already emphasized that I am no expert in strategic planning, and because this is not a book about strategic planning, this section will be short. My main objective with this section is to point out that the defined strategy of an organiza‑ tion is another ­high-­level element that contributes to setting the direction for the business process improvement efforts. The strategy, or rather the broad awareness of the strategy throughout the organization, should make sure that improvement projects are not allowed to go against it, as in the following example, or take off in any direction, in complete anarchy.

Example A ­medium-­sized European airline had traditionally catered primarily to the ­business-­travel segment, offering a wide range of destinations and depar-­ tures with high service levels. During the 1990s, the ­low-­cost competitors entered the market, typically first targeting the tourist market, but gradually also capturing increasing ­business-­travel market shares. With tougher price competition, the pressure for cost savings mounted, and several projects were initiated. Several changes were implemented over a couple of years: • The departure frequency to many destinations was significantly reduced • Several ­business-­class lounges were closed • Gate staffing was reduced, leading to a situation where business travelers could no longer go directly to the departure gate and change their booking there • Newspapers and meals included in the fare were eliminated or dras-­ tically reduced As a whole, the airline became much more like its ­low-­cost competitors, but still had somewhat higher prices. For the traditional business travel-­ ers, this made the airline a much less attractive supplier. There was a clear position in the top management that the business segment should still be a prioritized customer group, especially since it represented the most profit-­ able customers. In reality, through these ­so-­called improvement projects, many passengers were driven to choose other airlines that still offered ­business-­class service at traditional prices.

18  Chapter Two

Just as for the stakeholders, reviewing the organization’s defined strategic direction can be a worthwhile exercise, both to make sure the improvement projects are in line with it and to uncover areas where new improvement projects should be considered. Before moving on to a small selection of tools and techniques that can be used during strategy reviews, I should modify myself slightly. I might have given the impression that all improvement initiatives should be carefully assessed in terms of strategic fit before moving forward. Although I generally discourage too much freedom in starting projects locally without any coordination with other projects and priorities, there are exceptions. There are many examples of outsider or “ren‑ egade” projects that have altered or redefined the strategy of the organization. If you never allow any projects that fall outside the current strategy, you might very well lose such opportunities. This is very much in line with the famous work by Mintzberg (1994) about emergent strategies, a pattern of action that develops over time in an organization in the absence of a specific mission and goals, or despite a mission and goals. It is naïve to believe that all improvement efforts will be aligned with strategy, and perhaps they should not, either. Some of them might be break‑ through projects that ultimately shape a new strategy or create large gains. I will conclude by presenting a few tools and techniques that can be used when formulating and reviewing strategies. They are all established approaches, and some or all are probably well known to you already. If so, let this be a quick repetition of the following: • SWOT analysis • Competitive forces analysis • Strategy map

2.3  SWOT Analysis Identifying elements inside the organization and its surroundings within the per‑ spectives of strengths, weaknesses, opportunities, and threats (SWOT), the SWOT analysis is probably one of the ­best-­known simple strategic techniques that exists. Its purpose is simply to create an awareness of forces that will impact the orga‑ nization in the future. By understanding these forces and making them known throughout the organization, better strategic decisions can be made and the whole organization will be better prepared for future developments. Conducting a SWOT analysis is technically simple. It mainly entails brain‑ storming within the four perspectives, preferably in a group with different dis‑ ciplines and areas of the organization represented. “Technically simple” was deliberately used since I know from experience that performing a good SWOT analysis can be difficult, especially in terms of coming up with those elements that are relevant. This takes training.

Understanding the Organization’s Stakeholders and Strategic Direction  19

Stepwise, the approach is as follows:

1. Compose a team to undertake the analysis, drawing on different levels of competence within the organization. Consider supplementing the team with external representatives.



2. For each of the four analysis perspectives, brainstorm issues that seem relevant.



3. Compile the analysis results, using a simple table or matrix.



4. Discuss which of the issues identified is believed to have the strongest influence on the organization’s future.



5. Outline strategies or actions to deal with the findings.

Example A small town has seen its population drop steadily over the last few years, and its administration considers developing a new, fairly large residential area to counter this trend. The likely location of the area is adjacent to existing residential areas, where there is already an existing infrastructure in place, but it will require extending two current roads and converting parts of a park/green area used by many families in the town for recreational activities. To avoid charging ahead with the plans before truly understand-­ ing the possible consequences, the town council has asked that a SWOT analysis be performed, both for the town itself as a home to families and for the specific development plan. The results from the analysis are summarized in Figure 2.5.

Strengths

Weaknesses

Slow-paced town with high quality of life House prices lower than larger towns and cities in the area The new residential area will highly utilize the existing infrastructure

Few employment opportunities Long distance to larger town/city centers The new residential area will ruin a popular recreational area

Opportunities

Threats

Rumors about about a new road close by New inhabitants could lead to more industry

Figure 2.5  SWOT analysis results.

Aging population makes the town less attractive for young families Expected protests from environmental groups against the development plan

20  Chapter Two

The conclusion from the group was that additional employment oppor-­ tunities/business activity in the town would be important to develop to make sure that the new residential area became attractive. Furthermore, the possibility of replacement green areas to appease protests against the plan should be looked into.

2.4  Competitive Forces Analysis The SWOT analysis is a very broad analysis intended to identify any trends or developments that could affect the organization. A more specific analysis into the competitive position of the organization and its likely development can also be useful as a precursor to ensuing improvement projects. Porter developed one rec‑ ognized approach during the 1980s (for more information, see Porter’s books from 1980, 1985, and 1990 on competitiveness). The analysis is based on the competi‑ tive forces that influence an organization, as shown in Figure 2.6. Briefly explained, these five forces are: • Competitors that could claim market shares from the organization and other effects from competition, for example, price cuts or increased service level requirements • Buyers’ bargaining power to obtain better terms or switch suppliers, thus impacting the organization, as well as possible backward integration by cus‑ tomers to threaten the position of the organization • Suppliers’ bargaining power to increase prices or worsen terms for the orga‑ nization, as well as possible forward integration by suppliers to threaten the position of the organization

Threat of new entrants

Bargaining power of suppliers

Rivalry among existing competitors

Threat of substitute products or services

Figure 2.6  Five competitive forces.

Bargaining power of buyers

Understanding the Organization’s Stakeholders and Strategic Direction  21

• Threats of new entrants that can become competitors, depending on entry barriers, switching costs for customers, and so on • Threats of substitutes that could displace the organization’s products or ser‑ vices, depending on the match with current offerings, buyer switching pro‑ pensity and costs, and so on Some have argued that one force or dimension is missing in the model, namely, the power and influence of other stakeholders in the environment. These could be actors like regulatory bodies that often decide standards or requirements, banks and other creditors, shareholders, and so on. Whether five or six forces are included, the purpose of the analysis is to create an understanding inside the organization of possible threats that could materialize in the future and weaken the competitive position. The steps in the analysis are very similar to that of the SWOT analysis:

1. Compose a team to undertake the analysis, drawing on different levels of competence within the organization. Consider supplementing the team with external representatives.



2. For each category, brainstorm issues that seem relevant.



3. Discuss which of the issues identified is believed to have the strongest influence on the organization ahead.



4. Outline strategies or actions to deal with the findings.

Example A supplier of automotive parts had over the years built up a highly suc-­ cessful customer portfolio encompassing about half of the European car brands. The parts were brake components, made from brass, that were assembled into the brake system together with parts from many other sup-­ pliers. Recently, rumors were going around that other suppliers of simi-­ lar components were experimenting with composite materials. The use of composite materials would significantly reduce the weight of the parts, an important consideration for all carmakers. To decide whether to pursue such a strategy, a competitive analysis was performed. A small team was put together, and it decided to follow the five forces model. A selection of the more significant forces identified is presented.

Rivalry • Competitors that in general offer better terms to customers (con-­ tracts up for tender every third year)

22  Chapter Two

• Competitors that develop composite parts, thus offering lower weight (price consequences unknown) • Competitors that move toward becoming more comprehensive brake system suppliers, including the parts supplied by the company

Buyers’ Bargaining Power • Customers becoming even more aggressive in negotiations, demanding further price reductions during the life of the contract (the higher the risk, the more standard the product) • Customers awarding the company’s contract to competitors

Suppliers’ Bargaining Power • Few issues; most suppliers were local and depended on the suc-­ cess of the company

New Entrants • If the industry moves toward composite brake materials, suppliers with expertise in this material could target this market segment

Substitutes • Ongoing development work in many places in the world could lead to the successful introduction of electronic brake systems that could replace ­fluid-­based systems With all these forces on the table, the company realized that it had come to a turning point. Carmakers were constantly pressured to reduce emis-­ sions, and weight was playing an ­ever-­more important role, irrespective of the propulsion technology used. The likelihood of either plastic compos-­ ites or electronic systems replacing brass components in the near future was high, and there seemed no alternative to going down that road. As a consequence, a major improvement project was launched, with the aim of developing the first approved composite solution.

2.5  Strategy Map Developed by Kaplan and Norton (1996), the strategy map was briefly discussed in the groundbreaking book about the balanced scorecard. A more comprehensive treatment is found in the book dedicated to strategy maps (Kaplan and Norton, 2004). The concept of a strategy map is to force the organization to put words to both its end strategic goals and the road for getting there. Making use of the per‑

Understanding the Organization’s Stakeholders and Strategic Direction  23

spectives defined in the balanced scorecard concept—financial, customer, internal business processes, and learning and growth—an organization charting its strate‑ gic course should define milestones that must be achieved along the way to reach the final objectives. A generic strategy map is shown in Figure 2.7. The steps for developing a strategy map are as follows:

1. Compose a team to undertake the analysis, drawing on different levels of competence within the organization. Consider supplementing the team with external representatives.



2. Start by defining the overall, ­long-­term strategic objectives.



3. Go down one level in the map and identify which strategic goals must be achieved in the financial perspective to reach the overall goals.



4. Continue this way through the other perspectives, always asking what must be achieved to fulfill the overall goals and the goals of the above perspective.



5. Compile the strategy map based on the strategic objectives identified at each level, linking these together to form strategic paths.

Strategic subject

Strategic subject

Learning

Internal

Customer

Financial

Strategic subject

Strategic objectives

Figure 2.7  Generic strategy map.

Strategic subject

24  Chapter Two

A completed strategy map should be helpful to the organization, both to clarify for its members the direction to pursue and to derive more specific improvement efforts required.

Example A shipping company was experiencing trouble regarding its environmental performance. Large customers required reports about different environ-­ mental issues (fuel consumption, emissions, bilge water discharges, and so on), and they complained that the company didn’t have a clear enough pro-­ file in this area. There had even been warnings that unless things improved in future contract negotiations, contracts might be transferred to other sup-­ pliers with a more consistent approach to environmental management. This prompted the company to review its entire strategy, which up to now had very much been focused on low costs and providing reliable logistics services. With the aid of an external consultant, a ­small-­business development team sat down to design a strategy map that would espe-­ cially address the environmental aspects of the company’s development. The strategy map is shown in Figure 2.8.

Industry-best environmental performance

Financial

Premium price for environmental performance

Customer

Internal processes

Learning and competence

Environmental performance at no extra cost for the company

Customer proud of supplier

Consistent environmental procedures

Broad environmental training

Figure 2.8  Strategy map for the shipping company.

New technology

Technology development program

Understanding the Organization’s Stakeholders and Strategic Direction  25

As a result of the exercise, two specific improvement projects were launched: one to train almost every single employee in environmental issues and one to develop new environmental technology to help reduce fuel consumption and emissions from the ships.

As the examples have hopefully demonstrated, the organization’s strat‑ egy strongly impacts which improvement efforts are suitable at any given time. Although strategic planning might seem foreign to people working primarily with quality improvement, having an eye on this part of the organization is therefore useful. The next chapter will discuss how to ensure that you have a good under‑ standing of the organization’s current business processes before embarking on more specific improvement projects.

3 Understanding Your Current Business Processes

D

uring a lecture about business processes, a student asked me whether a company consists of just business processes or also has the traditional departments. At that time I had to answer: “Both, of course!” (Ten years later, for some organizations, the truth would in fact be that they consist only of business processes and have abandoned the old departments.) Viewing organizations as consisting of processes offers many advantages, and such an approach is also historically correct. This entire book deals with how to con‑ duct improvements and is based on improvement efforts directed at business processes. This chapter deals with the core concepts of business processes and how to understand them.

3.1  From Processes to Departments to Business Processes At some point in history, people began coordinating their actions to be able to perform more extensive or more complex tasks than what could be accom‑ plished alone. We must assume that such coordination, the first type of what might be termed organizations or enterprises, was initiated by a demand. For example, when a farmer hired helpers for tasks such as plowing, sowing, and reaping, it was because the work was too extensive for the farmer to manage on his own. The focus was on satisfying a demand for food by performing the necessary processes. In other words, the actions were process oriented. The workforce was not split into specialists in plowing, sowing, and harvesting; rather, the hired help represented a pool of hands that carried out the necessary tasks. Soon the extent and complexity of the tasks expanded further, and the num‑ ber of people in different coordinated enterprises increased. For as long as enterprises with more than a few employees have been in existence, they have

27

28  Chapter Three

probably been organized into departments. It was no longer feasible to main‑ tain a workforce where everybody performed each task. The tasks became so complex that the individual worker had to specialize. Therefore, the logical step was to form departments consisting of individuals with similar areas of expertise. This expanded into a tradition with a solid foothold in all types of organizations—commercial, public, and nonprofit. Until a few years ago, this way of organizing was highly dominant and had completely replaced the origi‑ nal way of aligning the organization more directly according to the work to be performed. Organizing people and work into departments provided, and still provides, some benefits: • People were allowed to specialize within their field of expertise, thus devel‑ oping a highly refined set of skills • Such departments could often more easily attract other specialists of the same training • Costs from centralizing various functions (for example, finance, personnel, maintenance) were lowered • The workplace was more secure; employees knew where they belonged and which tasks they were supposed to perform • The organizational structure was more clearly defined and could easily be drawn and presented Thus, we can safely claim that at least some modern organizations consist of both functional departments and horizontal business processes (I say “at least” since

Department/Function Purchasing Dept.

Manufacturing Dept.

Engineering Dept.

Finance Dept.

Inbound logistics

Processes

Order handling

Product development

Figure 3.1  The contradiction between vertical departments and horizontal processes.

Understanding Your Current Business Processes  29

some organizations, as already mentioned, have completely done away with depart‑ ments and have organized their employees inside processes). Figure 3.1 depicts a typical organization, with its vertical departments and horizontal processes that run through these departments (Andersen and Pettersen, 1996). Lately it has become obvious that this contradiction between mode of organization and tasks has created several problems. As soon as people are established within a square in an organizational chart with departments like those shown in Figure 3.1, it often seems as if the lines of this square become solid boundaries within which these individuals must remain. Communication across the borders is limited, and members of a department will perform only those tasks for which their department is respon‑ sible. Processes that run across departments and require contributions from many organizational units naturally suffer. Employees inside a department are told where their priorities should lie, and many of them end up focusing on sat‑ isfying their superiors. And this, as Jack Welch is rumored to have said, results in a ­less-­than-desired focus on the customer: “with their face to the boss and their ass to the customer.” Consider a process such as the one depicted in Figure 3.2. It’s a simple pro‑ cess that starts with a salesperson receiving an order from a customer and ends

Sales dept.

Order processing dept.

Accounting dept.

1. Salesperson takes order and sends to regional headquarters

2. Order is verified and checked for accuracy

3. Credit is checked

Stock control dept.

Logistics & distribution dept.

Production dept.

7. Order is shipped

6. Order is assembled

Figure 3.2  A simple business process.

4. Stock availability is checked

5. Order confirmation and details are sent to factory

30  Chapter Three

when the order has been shipped (or ultimately when the products have reached the customer safely). If executed in a normally performing organization, this process should take anywhere from a few hours to a few days. When this process is played out across a number of typical departments, it looks like the version shown in Figure 3.3. All of a sudden, the process has to traverse several departmental boundaries. For each such crossing, it somehow ends up in a priority queue at the next department. Research has shown that a good rule of thumb is that each such boundary crossing costs one week in lost time. In this process there are five crossings, easily adding five weeks of execution time—much longer than what it could have been in a streamlined organization. Each department seeks to maximize its influence and authority while optimiz‑ ing its performance level. The result is usually that the whole is far from being more than the sum of the individual parts, and in the worst case, far less. Each department suboptimizes within its area of responsibility, which in turn leads to conflicting objectives and actions between departments. The total performance level of the organization is naturally in line with this (Rolstadås, 1995).

Sales dept.

Order processing dept.

Accounting dept.

1. Salesperson takes order and sends to regional headquarters

2. Order is verified and checked for accuracy

3. Credit is checked

Stock control dept.

Logistics & distribution dept.

Production dept.

7. Order is shipped

6. Order is assembled

4. Stock availability is checked

5. Order confirmation and details are sent to factory

Figure 3.3  The same simple business process, now with department boundaries.

Understanding Your Current Business Processes  31

Example A company that manufactures elements for the building industry delivers regular stock goods, standard goods that are not stored, and special vari-­ ants. One of the company’s main problems has been that the demand for production capacity has been very unstable: sometimes it’s far above the capacity limit, and other times it’s well below. In periods of high demand, delivery times have dramatically increased, which is a problem in the market. In the company’s monitoring of the sales department, a central perfor-­ mance measure was fulfillment of the sales quota for each period. This led to a situation in which a salesperson, when the quota for a period had been met, put all later orders “in the drawer” for the next period. This way, the employee made sure he had a reserve that could help fill his next quota if sales should be slow. This, of course, contributed to the uneven demand for production capacity and the ­sometimes-­prolonged delivery times. At the same time, important information was withheld from the planning department that, in turn, could have given better production plans. The salesperson’s opti-­ mization of his own performance level, as it was being measured by the company, led to an overall performance level for the company that was far from the best achievable.

These types of problems are the basis for the last decade’s change from view‑ ing the company as a number of departments to focusing on the business processes being performed. Several issues make this a logical transition: • Every process has a customer, internal or external, and focusing on the pro‑ cess ensures better focus on the customer • The value creation with regard to the end product takes place in horizontal processes • By defining process boundaries and the customers and suppliers of the processes, better communication and ­well-­understood requirements can be achieved • By managing entire processes that run through many departments, rather than managing individual departments, the risk of suboptimization is reduced • By appointing process owners, who are responsible for the processes, the traditional fragmentation of responsibility often seen in a functional orga‑ nization is avoided

32  Chapter Three

• Managing processes provides a better foundation for controlling time and resources • Repetitive business processes represent the best opportunity for the organi‑ zation to relentlessly polish performance by continually looking for ways to improve them Studies have also been undertaken that attempt to document the results achieved by an orientation toward business processes in an organization. One interesting study reported by McCormack in 2001 was based on a 32-question survey given to 100 international and U.S. companies to determine any relationship between business process orientation (BPO) and the factors of: • Business performance • Interfunctional conflicts • Interdepartmental connectedness • Esprit de corps (team spirit) The data were analyzed using correlation analysis and regression analysis, generat‑ ing the following findings: • Companies with strong measures of BPO showed better overall business performance • Companies with strong measures of BPO tended to have better esprit de corps, better connectedness, and less interfunctional conflict The results of the analysis had strong significance levels and represent strong support for the benefits of business process orientation. What, then, is a business process?

3.2  Definition of a Business Process One basic definition of a process is: A logical series of related transactions that converts input to results or output. To separate a company’s processes from any other forms of processes, the word business has been added to form the term business process. A business process can be defined in the following way (Ericsson Quality Institute, 1993, 22):

Understanding Your Current Business Processes  33

• A chain of logically connected, repetitive activities that • utilizes the organization’s resources to • refine an object (physical or mental) • for the purpose of achieving specified and measurable results/products for • internal or external customers. In a similar vein, business process reengineering pioneers Hammer and Champy (1993, 35) defined a business process as “a collection of activities that takes one or more kinds of input and creates an output that is of value to the customer.” An emerging view might be along the lines of, “A number of roles collaborating and interacting to achieve a goal.” Irrespective of which definition one subscribes to, most authors emphasize the repetitiveness of business processes as an important characteristic. Although I have been unable to locate the reference, I know at some time I read statistics on the preparation time of a Big Mac. As far as I can recall, it had been reduced by several hundred percent since its introduction. The reduction most likely is a result of new technology in food preparation equipment and a careful tweaking of every step of the process until it comes close to what is humanly possible. If the process weren’t repeated so many times a day across the planet, there wouldn’t have been such a tremendous platform for this tweaking. The fact that the process will continue to be repeated for many years to come makes it worthwhile to invest in further improvements, and all the repetitions provide a constant test bed for experimentation. But not only processes that are repeated to this extreme will benefit from a careful design and ­ improvement-­oriented experimentation. Some years ago, we laid down paving in our driveway using stones of two by four inches. Although we don’t have a very big driveway, the stones amounted to about 14 tons. Subtracting all the preparatory work and sealing afterward, simply carrying the stones to the right spot, putting them down in the right pattern (we used gray and red stones), and knocking them tightly into position took me about four days. In the beginning, I simply grabbed a few stones, dropped them on the ground, and put them in their places, believing this would be a few hours’ work. When I realized how slow the process was, I started thinking how I could speed it up. I experimented with a few different approaches—sometimes carrying enough stones for a couple of rows, other times sorting the stones beforehand and distributing them along the edge. Although this kept my mind going, it wasn’t exactly rocket science. To make it more interesting, I started timing how long it took to lay down one row of stones. Astonishingly, on day two of the job, I went from 21 to 12 minutes. Had I just plodded on as I did in the beginning, the whole operation would probably have taken two extra days. For this little “enterprise” the savings were significant, even

34  Chapter Three

though I worked only four days. Think about what a difference it could make for a professional paving company if it could make similar optimizations. Encouraged by this success, I have tried to apply the same principle at later occasions, although (fortunately) I have not yet come across another task that took four days of repeating the same process. And every time, it is obvious that the effort bears fruit. Before painting 23 ­clip-­on window crossbars a few months ago, I spent a few minutes considering how to best design the process. I figured out which way would give me the best access, where to place the paint bucket to reach it easily, and in which order to paint the different squares to avoid having to change my grip too often. Tweaking the process further as I went along, I am sure I easily saved myself three to four hours of painful work spent mostly on my knees on a hard concrete garage floor. Ultimately, such optimizations are a combination of a design that is well thought through in advance of the process (sequence of activities, how to perform each step, which equipment to use, and so on) and controlled experimentation and tweaking of the process. In fact, processes that are carried out only a few times, perhaps even only once, will benefit from advance planning. For more complex tasks, this is in fact project planning, which is an established process. A word of warning is perhaps also warranted. I once interviewed a quality assurance manager at a large construc‑ tion project. During the interview, he proudly showed me binder after binder of procedures carefully outlined in flowcharts as well as spelled out in text (sometimes in ­if-­then statements). If this didn’t make me wonder whether this guy had taken things to the extreme, a remark toward the end of the interview did. He confessed that keeping up this high level of professionalism in process design almost required practicing outside of work. At home, he had several procedures for making fam‑ ily life more efficient, including one for “meal execution” and one for “bedtime preparations.” I couldn’t help wondering whether he had a procedure for what was to follow successful completion of the bedtime preparations! Another main point in the definition of a business process is that it has a cus‑ tomer, either external or internal. Thus, almost all activities within a company can be seen as a business process or part of a business process. What are the most cen‑ tral business processes, then?

3.3  Classification of Business Processes There are a number of ways of classifying business processes. Many leading com‑ panies, in terms of process orientation, have conducted thorough analyses of their own enterprises and compiled lists of their central processes. For example, both Xerox and IBM have made lists of a different number of processes. These lists are specially designed for the tasks carried out in these companies.

Understanding Your Current Business Processes  35

At the same time, different interest organizations have done a similar job, but on a more general basis. The purpose for these organizations has been to create lists that are general enough to be used by a large number of companies. Some con‑ tributors in this group are the International Benchmarking Clearinghouse (IBC), the European Foundation for Quality Management (EFQM), and the Automotive Industry Action Group (AIAG). Modeling enterprises and defining general sets of business processes has come to be a field of its own, occupying many academics. For example, researchers at the University of Plymouth have defined a hierarchy of business processes that spans five levels and is very extensive. The processes are divided into three main groups: operate, manage, and support. A somewhat simpler and more practically oriented approach can be found in the work performed in the Norwegian TOPP (Productivity Program of the Tech‑ nology Industry) program, a research program on productivity managed from NTNU (Norwegian University of Science and Technology) and its research partner SINTEF in Trondheim. To create a foundation for developing methods for ­self­assessment and benchmarking, a framework of business processes was developed. In line with the thinking of Porter’s value chain, the processes were divided into primary and support processes: • Primary processes are the central and ­value-­creating processes of the enter‑ prise. They run straight through the company, from receiving supplies from vendors to activities on the customer side. • Support processes are not ­value-­creating processes directly, but rather activ‑ ities needed to support the primary processes. They include activities like financial and personnel management. The symbolism of this view is shown in Figure 3.4. The primary processes rep‑ resent the core activities that ultimately generate income and allow a company to

Management

Maintenance

Planning Logistics

Purchasing Service delivery

Manufacturing

Finances

Figure 3.4  Primary and support processes.

HR

Invoicing

36  Chapter Three

survive. The importance of the support processes is, however, evident in the fact that the primary processes could not be executed without them. Together they form a necessary symbiosis. To clarify, or perhaps complicate, things further, some of the support processes can been extracted into a new class, termed development processes. These are processes that are supposed to take the value chain and its primary and support processes to a higher level of performance. Examples are product development, per‑ sonnel training, and supplier development. In the ENAPS (European Network for Advanced Performance Studies) project, funded by the European Commission, a generic set of processes was developed on the basis of these classes (see Figure 3.5, ENAPS, 1997). Here, the primary processes were named “business processes” and split into four main processes, each with subprocesses. The classes of secondary and development processes were grouped under the heading “secondary processes” and renamed “support” and “evolution.” There have been other attempts at classifying processes in alternative ways, and although these are not included here, they are probably quite accurate as well. The main point of these exercises is to create a ­high-­level understanding of which types of business processes exist in an organization.

ENAPS Generic Framework

Business processes

Secondary processes

Product development • Product research • Product engineering and design • Process engineering and design • Co-engineering

Support • Financial management • Human resource management • Information management • Maintenance • Internal control of health, environment, and safety

Obtaining customer commitment • Market development • Marketing and sales • Tendering Order fulfillment • Procurement and inbound logistics • Production planning & control • Manufacturing and assembly • Distribution and outbound logistics • Order processing Customer service • After-sales service • Product take-back

Figure 3.5  Business processes in ENAPS.

Evolution • Continuous business process improvement • Product research • Production technology research • Human resource development • Supplier base development • Development of external relations • Strategic planning

Understanding Your Current Business Processes  37

3.4  Identifying the Organization’s Business Processes Before documenting one or more business processes, a prerequisite is that the business processes of the enterprise have been identified. This can at times be rather difficult, as it is rarely obvious which processes are undertaken by different departments in a functional organization. Two complementary routes to such an understanding can be utilized (Peppard, 1998). The most direct way is to simply generate a list of the business processes that an organization is believed to encom‑ pass. Such a job will often be based on existing process descriptions or procedures written for ISO 9000 certification or similar purposes. A more rewarding and systematic route is to map out the following: • The strategy of the organization • Its stakeholders, that is, organizations, institutions, or people affected by or with a vested interest in the organization and its business processes • Expectations with regard to products or services delivered by the organization • The business processes that produce these products or services and support and enable the delivery of them The relationships between these elements are shown in Figure 3.6. Hopefully, you realized that the first three items are the same elements that have been discussed so far in the book. By going through this list, it is much easier to point out the business processes that are or should be carried out by the orga‑ nization and that are necessary in fulfilling the expectations of its stakeholders. Ideally, your organization should have exactly those business processes that are required to satisfy the strategy defined and its stakeholders. Anything beyond these could be seen as waste, unless they are support processes necessary for ­performing Strategy shape

encapsulated in defines

delivers

Stakeholders hold

Processes deliver

satisfy

consolidate

Expectations Figure 3.6  Business processes derived from strategy and stakeholders.

38  Chapter Three

the primary processes; anything less could lead to dissatisfaction on the part of some stakeholders. By nesting backward from the output required by the stakeholders to the pri‑ mary and support processes and input into these, several strings of business pro‑ cesses emerge. Even if this approach does not cover every conceivable process ever performed by the organization, this is OK. Processes not encountered when backtracking from the stakeholders’ expectations are hardly crucial in satisfying the stakeholders. Thus, if omitted, they will rarely be missed.

3.5  Business Process Modeling A general rule for improving something is that it is necessary to know the current state of things. This is very true for business processes as well. If you do not know how the process looks and works today, it will be very difficult to know which improvement initiatives can be started and whether they will work. Documenting one’s own process should therefore always be the first step in any improvement activity. Modeling, in general, carries with it several other advantages: • Modeling requires confrontation and clarification where disagreements or ambiguities prevail, and it does not allow vague assumptions. • Modeling leads to a general increase in the level of specification; assump‑ tions, surroundings, and objectives must be explicitly expressed. • A model also allows more advanced simulations or scenario testing and can be used as a tool in ­what-­if assessments. • A business process model represents a documentation of deliberately designed processes. It can be used to demonstrate to customers, regulatory bodies, or other actors that the organization employs a systematic approach to process design and fulfills requirements posed. For customers, especially industrial or potential, the model can be used as a marketing tool to con‑ vince them that your business processes are sound and reliable. • Many types of standards and certification require extensive process model‑ ing, including the ­process-­based year 2000 version of ISO 9000. • If modeling is done right, that is, in collaboration among many or all of those involved in the business process being modeled, it stimulates involve‑ ment and enthusiasm. • The dialog connected with the modeling process creates learning and a better understanding of the process and its boundaries. This might be the most impor‑ tant benefit; after an extensive business process modeling effort in an organiza‑

Understanding Your Current Business Processes  39

tion, employees are much more aware of their role in the bigger scheme of things and how their effort contributes to the overall objectives of the organization. Business processes can be documented at two different times: • One by one in connection with improvement projects involving the specific process • All at once at the start of a general “improvement journey” You might select an approach where a process is not documented until a project or some other activity is started to improve it. In such cases, process documentation is the first activity of the project and serves several purposes: • A common understanding within the improvement team of the content of the process is created, including its activities, results, and who performs the different steps • The scope of the process is defined, as are the boundaries to adjacent processes • If desired, it is possible to define more specific problems within the main process This approach works quite well in practice. There is no need to document many processes at the start, as this is carried out when the need arises. For smaller organi‑ zations lacking resources or for organizations where the business processes change rapidly, this is probably the preferred approach. The other way is to initiate the improvement journey by going through the entire organization and modeling all or most of the business processes. As mentioned in sec‑ tion 3.3, this might involve everything from 15 to 100 business processes. It is not pos‑ sible to cover every single process performed within the enterprise; the purpose is to document the most important ones. The advantages of this approach are as follows: • By involving a large number of employees in this work, a positive atti‑ tude and improvement motivation that are useful in later projects are often created • After undertaking such an exercise, top management will have formed a very good overview of the organization and the need for specific improve‑ ment projects • Insight into the individual processes that is created during the documenta‑ tion work often makes it possible to pinpoint elements of individual pro‑ cesses that can be improved

40  Chapter Three

To the extent that the time and resources necessary for such an introductory pro‑ cess modeling can be mustered, this is probably the most valuable approach. Many organizations today have done this exercise of modeling for most of their processes. For many, it has been part of the path toward ISO 9000 certification; for others, it’s simply a useful analysis. However, bear in mind that process models and pro‑ cedures in quality handbooks often reflect an ideal situation, not always the real situation. In improvement projects, it is essential to start from the real situation to be able to identify problem areas. Furthermore, most processes are to some extent dynamic, often rendering existing descriptions inaccurate or erroneous. Checking the quality of such material before using it can save much extra work. The objec‑ tive must always be to document the process as it is performed today, not the way it was supposed to be performed. Before moving on to more specific techniques for modeling business pro‑ cesses, I should point out that it might not be as straightforward as I have perhaps made it seem so far. There is, in fact, a hierarchy of models at different levels that could be employed, as shown in Figure 3.7 (which I should point out represents the most complex situation; you might say a worst case). Let me quickly introduce each of these: • The stakeholder model has already been discussed (see section 2.1), and it is merely a graphical representation of the stakeholder analysis. In this hierarchy, it gives the ­highest-­level view of the organization. • The industrial value system, often termed value chain, more specifically defines the organization’s position in the larger chain of ­supplier-­customer relationships that form the chain, from raw material suppliers to end ven‑ dors of products or services to the final customer. Every organization is part of one or more such value systems, and understanding your position in them is an important insight to bring along in the process modeling task. I will not pursue the topic of value systems/value chains/supply chains any further in this book, as it is simply too comprehensive to allow a fair treatment here. • The overall process model is perhaps the first true business process model in the hierarchy, but it does not represent an individual process. Rather, such overall models attempt to illustrate how the individual processes are linked, in chains or networks, inside the organization. They can often make the big picture clearer than if you jump directly to single processes. There are many ways to construct an overall model; the one shown in the figure merely offers an example. Some models of this type are quite detailed, indicating flows that exist between the different processes; others simply portray them together in the same chart. I will not discuss this level of modeling in much more detail here, but you should decide whether you will include this level of modeling and, if so, in which style. How to identify the business processes that should go into the model, however, was briefly discussed in section 3.4.

Understanding Your Current Business Processes  41

Stakeholder analysis Authorities

Owners

Management

Customers

Suppliers Financial institutions

Competition Alliance partners

Environment

Employees

Industrial value system A

B

C

D

Overall process model Primary processes

Marketing and sales

Order processing

Support processes

HR management

Development processes

Product development

Purchasing

Manufacturing

Maintenance

Market development

Distribution

Invoicing

Safety, health, and environment

Supplier development

Relationship map Relevant process

Supplier

Process X

Quality check and agreement approval

Negotiations

Process X

Quality control

Information about price/offer

Identify suppliers, negotiate and sign agreement

Request for price/offer

Need for new or revised agreement

Agreement evaluation

Request offers and perform negotiations

Signed agreement

Negotiations for articles for which no agreement exists

Supplier

No

Process X

Defined article?

Yes

Place order with supplier according to agreement

Follow-up of order

Order

Follow-up and response

Supplier

Supplier

Receipt of product or service

Payment

Invoice

Payment for delivery

Supplier

(continued)

Figure 3.7  The hierarchy of process modeling. Individual flowcharts Marketing department

R&D department

Product planning

Manufacturing

article?

Yes

or service

to agreement Order

Follow-up and response

Supplier

Supplier

Invoice

Payment for delivery

42  Chapter Three Supplier

Individual flowcharts Marketing department

R&D department

Product planning

Manufacturing

Process planning

Trial manufacturing

Initiation of new-product development

Design of product concept and suggestion for technical solution

Development of specifications

Process certification

Product ready for manufacturing

Procedure xx 1. Step a 2. Step b 3. Step c 4. Step d 5. Step e 6. Step f 7. Step g

Figure 3.7  Continued.

• In between the overall process model and the individual flowcharts, some find it helpful to include a relationship map. This portrays the main inter‑ faces and linkages among units inside the organization itself and external actors that interact to perform the business process. • Individual flowcharts, which are the most common type of process models and the ones that are most useful in process improvement work. There are several different styles of modeling, and these will be discussed in more detail shortly. • Finally, I have also included the type of ­text-­based written procedures that were more common some years ago. When developed into a high level of

Understanding Your Current Business Processes  43

detail, these are probably the most precise process descriptions. For example, they often include detailed instructions about which button to press in a com‑ puter program or how to hold a mechanical tool. However, they are often difficult to understand quickly, and they rarely contribute to the overall under‑ standing of a process. Thus, I generally do not worry too much if this level is not covered when the aim of the modeling is improvement purposes. The following sections present in more detail some of the most commonly used and, in my opinion, most useful modeling approaches: • Relationship mapping • Traditional flowchart, including the IDEF0 standard • Cross-functional flowchart • Flowchart divided into process segments • Several-leveled flowchart • Flowchart with statistics

3.6  Relationship Mapping Before we can draw a detailed flowchart of a process, it is often necessary to create an overall picture of the individuals who are part of the process and their relation‑ ships with one another and the rest of the world. This is especially true for more extensive and more complicated processes involving a number of individuals or departments. Is the objective, for instance, to document the process of order receipt and delivery of goods to the customer? If so, it can be quite a challenge to put in place the separate steps of the process just like that. In such a situation, relationship mapping is an ideal first step. In contrast to an ordinary flowchart, a relationship map does not consider activities or the sequence of such. The map is constructed by placing on a blank sheet the different units, department, or individuals expected to participate in or impact the process. For the process of order receipt and delivery, logical partici‑ pants would be the sales, planning, production, and procurement departments internally, as well as customers and suppliers. Furthermore, we might assume that the finance department and external transport companies would be included. A general rule is that it is better to include too many elements, as irrelevant ones will naturally be eliminated throughout the process. It is also possible to draw such a map on several levels so that each department can be further detailed on a lower level without blurring the overall picture.

44  Chapter Three

After establishing the potential participants in the process, each relationship among them is analyzed to define its type. Different types of arrows are suitable for this purpose. Elements without any relationships to other elements are removed from the map. In the end, the map is redrawn and will give a good overview of relationships between participants and stakeholders in the process. Figure 3.8 shows a relationship map for the sample process. The types of arrows used in this map are only suggestions; there are no standards in this area. It should also be pointed out that the task of conducting such a relationship mapping and the preceding activities of constructing a flowchart and other tasks related to process documentation must be carried out in a group that includes the most central participants in the process. The objective is to improve and adjust the process model until it reflects how the process is performed today. On the other hand, you should be careful not to spend too much time creating a very detailed and accurate depiction of the process. It might very well be that a reasonably good model achieved quite easily is better than one that demands many resources to generate but is perfect. This can be seen in the context of the later use of the process model. To summarize, the procedure for constructing a relationship map is:

1. Identify the internal and external organizational units that play a part in the business process.



2. Construct a diagram where these units are placed in logical positions toward one another.

Own company

Planning

Manufacturing

Purchasing

Customers

Sales

Suppliers Flow of goods

Information

Order

Discussion/negotiations

Figure 3.8  Example of a relationship map.

Understanding Your Current Business Processes  45



3. Go through the entire diagram and consider if there is a relationship between any two units. For each relationship identified, indicate this in the diagram using a type of arrow that corresponds to the type of link that exists.



4. After completing this analysis, remove from the diagram any units that do not have links to other units, unless it is a point to demonstrate specifically that no links exist.



5. If the diagram becomes difficult to interpret, rearrange it by moving units with extensive links closer to each other.

Example A large international corporation was organized with one central manu-­ facturing site covering all of Europe and local dealerships with their own inventories of finished goods in a number of countries. There were clear indications that the supply process, including communicating needs from the local dealers and distribution to them, was not working satisfactorily. Thus, the company started a project to improve the material flow, begin-­ ning with the main manufacturing site. However, it proved difficult to form an overall impression of the flow of information and goods, and it was thus decided to start the project by drawing a relationship map. The first step was to collect information from all involved parties about the most important transactions they were involved in: • The local dealers based their demand prognoses on ­ twice-­yearly sales consultations with their main customers • Using the aggregated information from these consultations, the dealers transmitted their expected demand for the different product families to the manufacturing unit • Here, information from all the dealers was further aggregated into rough prognoses for the next six months • These formed the basis for negotiations with suppliers about frame agreements for the following period • The local dealers issued detailed orders every month with a fixed delivery time from the manufacturing site of three weeks • The ordered products were either taken from a small ­finished-­goods inventory or manufactured before they were sent by truck to the dealers The local dealers were invoiced only when the goods had been sold to end customers. This produced the relationship map shown in Figure 3.9.

46  Chapter Three

Suppliers Negotiations about frame agreements

Central manufacturing site

Goods

Local dealers Goods

Goods

Prognoses and orders

Goods

Prognoses and orders Goods

Sales consultations

End End customers customers

Sales consultations

Local dealers

Sales consultations

Local dealers Goods

End End customers customers

Goods

End End customers customers Figure 3.9  The resulting relationship map for the supply process.

3.7  Traditional Flowchart A flowchart is a graphic depiction of the flow of activities in a process. The use of flowcharts reinforces the fact that it is much easier to understand something presented graphically instead of described by words. Put differently, an image is worth a thousand words. There are many ways of drawing a flowchart (Andersen and Pettersen, 1996). The most basic is simply to use different symbols to represent activities, and arrows to illustrate the connections between the activities. There are a number of variants in the symbols used, including everything from complex pictures to simple boxes and lines. One way is not necessarily better than another; the point is simply that the users must understand the symbols. Some commonly used symbols are: Start or finish point Step or activity in the process Decision point Input or output Document

Start or finish point Step or activity in the process

Understanding Your Current Business Processes  47 Decision point Input or output Document Manual operation

Storage or waiting

In addition, it is possible to indicate on the side of the symbols in the flowchart what equipment or resources are being used and under what conditions the activ‑ ity is being performed. A flowchart for the sample process discussed earlier might look like the one shown in Figure 3.10. This chart could have been made more detailed—for instance, by including the suppliers in the process, the negotiations with the suppliers and customers, and so on. Nevertheless, it illustrates the principles for drawing a flowchart.

Customer demand Ordering Customer order

Production planning Production plan

Procurement needs

Procurement

Manufacturing

Products Delivery Satisfied customer

Figure 3.10  Flowchart for the supply process.

Purchased parts

48  Chapter Three

The steps to design a flowchart are as follows:

1. Define the boundaries for the business process to be modeled, especially the start and end points. Both should be ­well-­defined outputs or events.



2. Starting from the end point, identify the activities carried out in the process, along with important outputs generated, periods of waiting, and so on.



3. Construct the flowchart graphically by placing items in correct sequence, moving from the end of the process backward. In cases of decision points or places where the process branches out, try to orient the diagram so that activities flow in the most suitable direction, that is, from left to right and top to bottom.



4. In cases of disagreement in the modeling team, take time to resolve any issues so that the final model truly reflects a common understanding. Remember that such disagreements often represent good opportunities for clarifying potential conflicts or misunderstandings.



5. When the team is satisfied that the most important elements of the process have been captured and placed in the correct sequence, it is a good idea to redraw the flowchart to increase readability.

Example A group of secretaries in a public office experienced problems with docu-­ ments and other materials that, once they were filed, were difficult to find again. Among the employees, there was a hunch that some people used dif-­ ferent methods for filing than others. It was therefore decided that the way the process is carried out—and how it really should be carried out—should be defined exactly. For this purpose, they wanted to use a flowchart. Armed with a white board and pads of small yellow stickers, the sec-­ retaries gathered in a meeting room. It was soon evident that they all per-­ formed approximately the same steps when filing documents, but the criteria for where material was filed varied quite extensively among them. After a heated debate, they were able to agree on both the process and the filing criteria. The flowchart in Figure 3.11 is the result of the exercise.

Understanding Your Current Business Processes  49

Material for filing

Decisions and resolutions

Type of material?

Letters, faxes, and e-mails

Complaints File under “closed cases”

File under “cases under processing”

File under “correspondence”

Criterion alphabetically

Criterion date

Criterion case

Update document list Filed material

Figure 3.11  Flowchart for the filing process.

When presenting the tool of flowcharting, I would be negligent not to men‑ tion IDEF0, often referred to as IDEFØ, that is, ­IDEF-­zero. IDEF0 (where IDEF is an acronym for Integrated DEFinition method) is best described as a standard for modeling decisions, actions, and activities of an organization or system. It was derived from a ­well-­established graphical language called the Structured Analysis and Design Technique (SADT). Because of its consistency in its notation and rules for how flowcharts are designed, it is quite a powerful approach. However, I find that the resulting flowcharts, especially for more complex processes, become quite difficult to read. They are correct in the sense that all flows are captured, but seeing the big picture can often be difficult. Therefore, I think IDEF0 is best suited for ­high-­level modeling of a process. In an IDEF0 diagram, four types of links between items in the chart are defined: • Controls: requirements, laws, company policy, and so on, that dictate the behavior of the step in the process • Mechanism: resources that carry out or support the process

50  Chapter Three

• Input: information, materials, and so on, that initiate the process • Output: the result of the process (transformation of input) These four connections are graphically linked to one predetermined side of the square representing the process step, as shown in Figure 3.12. Figure 3.13 shows a typical IDEF0 flowchart, and I’d like to point out that this is a fairly simple process that has been modeled. Even so, you can already

Controls

Process

Input

Output

Mechanism

Figure 3.12  Standard IDEF0 diagram design.

RC milestones C2

Technical annex I1

Management & control (internal, EC, audits) C1

Evaluation feedback C3 & recommendations

Market C4 feedback

WP R1 Define reqs + specs

Reqs specs 03

P. 5 LC models V1

WP R2 Develop PROMISE model

Info-flow models v1

A112 P. 6 WP R3 Define AS & demos A113 P. 7 External Projects M1

Figure 3.13  Sample IDEF0 diagram.

M2 Partners

Applied scenarios 01 02 Demo definition

Models 04

Understanding Your Current Business Processes  51

see that the diagram has become quite complicated. Compared with the simpler, albeit perhaps technically not as correct, flowchart shown in Figure 3.10, this one is more difficult to relate to and understand the logic of. Although I might sound overly critical toward the IDEF0 standard, I have found the approach less useful in ­real-­life business process modeling. The notation of IDEF0 in Figure 3.12 is useful for guiding the modeling effort. Other notations also exist, such as the turtle approach. The turtle diagram was designed to remind people who are modeling business processes to capture the fol‑ lowing relevant information: inputs, outputs, equipment required, skills required, work instructions, and performance measures. All such standards have been designed to ensure that the flowchart is logi‑ cally constructed. One problem with standard, simple flowcharts is that it can be difficult to see who performs which tasks. This information is better presented in ­cross-­functional flowcharts, which is the next topic.

3.8 ­ Cross-­Functional Flowchart As has been mentioned, an ordinary flowchart mainly describes what activities are performed in a process. A cross-functional flowchart shows who performs the activities and which functional department they belong to—from which the name arises (Andersen and Pettersen, 1996). Figure 3.14 shows an example, where the ordinary flowchart in Figure 3.10 has been supplemented with more detailed infor‑ mation. Each department has been placed inside a “swim lane,” and activities per‑ formed by the department are located in this column. Adding these pieces of information does not take much time compared with the job of describing the sequence of activities itself, but it contributes to a much clearer representation of the process. It is therefore generally recommended to use a ­ cross-­functional flowchart. Another advantage of this type of flowchart is that it allows you to analyze when responsibilities in the process shift. As was discussed earlier in this chapter and shown in Figure 3.3, when a business process traverses a department boundary, delays in the process are usually introduced. A ­cross-­functional flowchart makes it easy to see how many times and where this happens, thus allowing for improvements by either redefining responsibilities or placing steps performed by the same department in sequence. Frequent shifts between departments or actors in the process also indicate a danger that the overall process responsibility is highly fragmented, increasing the likelihood that no one will take charge for the performance of the process. I should also mention that the swim lanes can denote external players in the process, physical locations where steps of the process are executed, or other units. It is also possible in a ­cross-­functional flowchart to display additional information. Along the vertical axis, or the horizontal axis if the process is drawn in a landscape format, information such as the following can be added to the chart:

52  Chapter Three

Customer

Planning department

Procurement department

Manufacturing department

Distribution department

Customer need Ordering Production planning Procurement Manufacturing Delivery Satisfied customer

Figure 3.14  Example of a ­cross-­functional flowchart.

• Time spent so far in the process • Incurred costs thus far in the process • Value added • Degree of completion With these additional details, the flowchart is able to convey much more informa‑ tion than just the pure sequence of activities in the process. To construct a ­cross­functional flowchart, proceed as follows:

1. Define the boundaries for the business process to be modeled, especially the start and end points. Both should be ­well-­defined outputs or events.



2. Starting from the end point, identify the activities carried out in the process, along with important outputs generated, periods of waiting, and so on.



3. For each of the items in the process, determine which organizational unit is in charge of executing the item.



4. Construct the flowchart graphically by creating swim lanes or columns that correspond with the organizational units involved, preferably in such a sequence that units with close cooperation are located next to each other.



5. Continue by placing items in the correct sequence and in the correct swim lane, moving from the end of the process backward. In cases of decision

Understanding Your Current Business Processes  53

points or places where the process branches out, try to orient the diagram so that activities flow in the most suitable direction, that is, from left to right and top to bottom.

6. In cases of disagreement in the modeling team, take time to resolve any issues so that the final model truly reflects a common understanding. Remember that such disagreements often represent good opportunities for clarifying potential conflicts or misunderstandings.



7. When the team is satisfied that the most important elements of the process have been captured and placed in the correct sequence, it is a good idea to redraw the flowchart to increase readability.

Example A large company had noticed during the last few years that it was quite expensive to generate the necessary information for financial reporting to var-­ ious public institutions, including tax authorities. To gain an overview of who did what in the process and where costs were incurred, a ­cross-­functional flowchart was constructed. The resulting chart, shown in Figure 3.15, was well suited for this purpose and helped reduce costs considerably through a redistribution of responsibilities for different tasks. Effort was made to ensure

Incurred costs

Authorities

0

Report call

Financial manager

100

Assessment of call

250

Identification of necessary information

Clerk

Collection of information

600

Submission of information

3900 Compiling of information

4300 Report generation

4800

4800

All departments

Received report

Figure 3.15  ­Cross-­functional flowchart for the reporting process.

54  Chapter Three

that the necessary information from various departments was systematically collected, rendering infrequent, ­all-­out efforts unnecessary.

Please keep in mind that the more information that is added, the more com‑ plex the flowchart typically becomes. Dividing it into logical areas is one way of increasing the readability.

3.9  Flowchart Divided into Process Areas Even flowcharts not designed using the IDEF0 notation can become too complex to be easy to work with. The chart in Figure 3.14 is perhaps not a good example, but if the process to start with is long and complex, or much other information is added to the flowchart, it might become difficult to read. Two possible ways of dealing with this are (1) divide the flowchart into distinct process areas and (2) split the flowchart into several smaller charts (the second approach is dealt with in section 3.10). Dividing the chart into process areas is simply a matter of identifying logical segments of the process that can be grouped together. These can then be set apart from other areas by drawing them in different colors, placing them inside rect‑ angles with borders separating them from other process areas, or shading the area of the process. An example of this is shown in Figure 3.16. Although this is not a complex process in itself, it illustrates the principle. This is a process consisting of three distinct segments: sea transport, coordination of land transport, and the actual land transport. By placing these inside three separate rectangles, they are logically kept apart, and each can be viewed separately within the big picture. For processes with tens, if not hundreds, of steps, this is quite effective. Since this is simply a variant of the traditional flowchart, a more detailed example of its use has not been included. The steps to create such a chart are simi‑ lar to those of a regular flowchart:

1. Define the boundaries for the business process to be modeled, especially the start and end points. Both should be ­well-­defined outputs or events.



2. Starting from the end point, identify the activities carried out in the process, along with important outputs generated, periods of waiting, and so on.



3. Construct the flowchart graphically by placing items incorrect sequence, mov‑ ing from the end of the process backward. In cases of decision points or places where the process branches out, try to orient the diagram so that activities flow in the most suitable direction, that is, from left to right and top to bottom.

Understanding Your Current Business Processes  55

Land logistics coordinator

Land transporter

Finished vehicles logistics

National autologistics

Follow up status and invoices for transshipment

Land transport status

Distribution network

Deep sea transporter

Shipment details

Freight lines Send vessel schedules Timer

Transport contracts handling

Create and negotiate transport routes

Book transport Land transport instructions/ volume forecast

Figure 3.16  Example of a flowchart divided into process areas.



4. In cases of disagreement in the modeling team, take time to resolve any issues so that the final model truly reflects a common understanding. Remember that such disagreements often represent good opportunities for clarifying potential conflicts or misunderstandings.



5. At this point, examine the flowchart for any distinct phases or areas of the process. If useful, these can be separated into their own areas, using either shading or rectangles to create boundaries around them.



6. When the team is satisfied that the most important elements of the process have been captured and placed in the correct sequence, it is a good idea to redraw the flowchart to increase readability.

3.10 ­ Several-­Leveled Flowchart Flowchart complexity can also be remedied by breaking the chart into several hier‑ archical levels, for both ordinary and ­ cross-­functional flowcharts (Andersen and Pettersen, 1996). The principle is that the top level shows only the main activities, which are numbered 1.0, 2.0, and so on. This top level gives a good overview of the entire process without blurring the major aspects with many details. Figure 3.17 shows a somewhat simplified version of the chart in Figure 3.14 as a level 0 chart. To present the main activities in more detail, another flowchart on the level below this one is created. For activity 3.0, the more detailed chart is shown in

56  Chapter Three

Customer demand

Ordering

Production planning

Procurement

1.0

2.0

3.0

Manufacturing

Delivery

4.0

5.0

Satisfied customer

Figure 3.17  Level 0 chart for the process.

Figure 3.18. The individual steps at this level are logically linked to the level above by numbering them 3.1, 3.2, and so on. If there is a need for further detail, more levels are simply added. With this technique, it is possible to pre­ sent the overall process using only a level 0 chart. As the need occurs, details are displayed for the individual steps of the process through ­lower-­level charts. This makes the technique very powerful through combining a lucid presenta‑ tion format with details where needed. If the flowcharts are stored and accessed electronically, perhaps as part of an intranet accessed through a Web browser, the hierarchical representation is well suited for navigation down through the levels. Simply clicking on a step in a chart on a higher level takes you to the more detailed chart for that step. A set of flowcharts at different levels is created through the following steps:

1. Define the boundaries for the business process to be modeled, especially the start and end points. Both should be ­well-­defined outputs or events.



2. Starting from the end point, identify the main, ­ high-­level activities car‑ ried out in the process, along with important outputs generated, periods of waiting, and so on.



3. Construct the flowchart graphically by placing these ­high-­level items in correct sequence, moving from the end of the process backward. In cases of decision points or places where the process branches out, try to orient

Understanding Your Current Business Processes  57

Procurement list

Selecting supplier

Negotiating price

Placing an order

3.1

3.2

3.3

3.4

Receiving order confirmation

Registering a delivered order

3.5

Parts for manufacturing

Figure 3.18  Level 1 chart for procurement.

the diagram so that activities flow in the most suitable direction, that is, from left to right and top to bottom. Activities in this ­upper-­level chart are numbered in sequence from start to finish.

4. In cases of disagreement in the modeling team, take time to resolve any issues so that the final model truly reflects a common understanding. Remember that such disagreements often represent good opportunities for clarifying potential conflicts or misunderstandings.



5. At this point, examine the flowchart for any ­high-­level activities that would benefit from having details presented in a separate flowchart. For each of these, separate charts are created by following steps 1–4. These are num‑ bered in such a way that they link the activities to the correct step in the ­upper-­level chart. If required, further flowcharts of even more detail can be created.



6. When the team is satisfied that the most important elements of the process have been captured and placed in the correct sequence, it is a good idea to redraw one or more of the charts to increase readability.

58  Chapter Three

Example To ensure that all departments used the same procedure for storing the necessary information for financial reporting, the company from the last example decided to document the ideal information flow of the process by using ­several-­leveled flowcharts. The flowchart in Figure 3.15 constituted the level 0 chart, and the step that represented the activity taking place in all departments was numbered 4.0. Because the reporting required dif-­ ferent information from different departments, ­department-­specific proce-­ dures were produced at level 1. For the procurement department, the chart is shown in Figure 3.19. 4.1

Monthly registering of: • Procurement value • Supplier distribution • Country of origin distribution • Delivery times

Quarterly registering of: • Spent workforce hours • Costs for wages • Value-added tax paid

4.2

4.3

Year-end registering of: • Total purchasing value • Supplier distribution • Country of origin distribution

All reports are generated and stored in spreadsheets

4.4

Figure 3.19  Level 1 chart for reporting.

In cases where processes branch out into several possible parallel paths, and where it is useful to understand the importance of each path, statistics can be added to the flowchart.

3.11  Flowchart with Statistics One problem I have encountered many times when modeling processes is that obscure exceptions or special cases require the same, if not more, attention than

Understanding Your Current Business Processes  59

the most common flow of events. Very often, you will find that a certain process is performed according to a standard sequence. For some customers or some prod‑ ucts, or due to other exceptions, there will be smaller or larger deviations from this norm. When modeling this process, the modeling team will insist on including the exceptions. This makes the flowchart more complex than what is strictly necessary. Some‑ times you’d be better off simply modeling the main flow and disregarding these exceptions, as this will be sufficient for understanding the logic of the process. In other cases, the exceptions might be what cause problems in the process, and they must be captured. In any case, adding some sort of load statistics or weights to the different flow possibilities can be helpful. These indicate how often the main and alternative flows are used. The approach is best demonstrated using an example, but let me first briefly list the steps to use when making such a chart:

1. Define the boundaries for the business process to be modeled, especially the start and end points. Both should be ­well-­defined outputs or events. Keep in mind that this type of chart is best suited for processes where there are two or more alternative flows.



2. Starting from the end point, identify the activities carried out in the pro‑ cess, along with important outputs generated, periods of waiting, and so on.



3. Construct the flowchart graphically by placing items in correct sequence, moving from the end of the process backward. In cases of decision points or places where the process branches out, try to orient the diagram so that activities flow in the most suitable direction, that is, from left to right and top to bottom.



4. In cases of disagreement in the modeling team, take time to resolve any issues so that the final model truly reflects a common understanding. Remember that such disagreements often represent good opportunities for clarifying potential conflicts or misunderstandings.



5. At this point, determine where in the flowchart the statistics could be use‑ ful, typically for parallel and alternative flows through the process. Use existing statistics or collect the required statistics about usage of the alter‑ native flows. The statistics are inserted into the chart to indicate how often each alternative is used.



6. When the team is satisfied that the most important elements of the process have been captured and placed in the correct sequence, it is a good idea to redraw the flowchart to increase readability.

60  Chapter Three

Example A small branch office of a bank had started a project to redecorate the offices. To determine the best layout of the refurbished office, it was impor-­ tant to understand the flow of customers in the bank. The main questions discussed were how much capacity, and thus space, to allocate to the main reception, which would be the first point of contact for customers, and how to structure the back offices, where customer representatives specializing in various areas would be located. During one week of service with the current office layout, all customer activity was logged. A flowchart depicting the physical flow of the custom-­ ers through the bank was constructed, and data showing the number of customers taking the different routes were added. The resulting chart is shown in Figure 3.20, and it helped the bank decide to increase the capac-­ ity (and level of experience) of the front reception. Many people visiting the bank had appointments, and they were simply shown into the offices of the people they came to see. Additionally, many had requests that could be settled in the reception area. If the requests could not be handled there, the customers were referred to the customer representatives, taking up valu-­ able time that the representatives could have used to pursue sales leads. Example continued

In concluding the treatment of business process modeling, I should proba‑ bly mention that other means of modeling processes have come into use lately. Some find that simple symbols are less intuitive in flowcharts. To make process charts more readable, using more detailed drawings or even photographs can help. Photos of actual machines, people, or locations can help liven up any flowchart. Some organizations have successfully used video to document processes. It’s especially helpful for those processes where a very precise sequence of opera‑ tions or performing steps exactly the right way is important. For such processes, video can offer much more detail than any flowchart or written procedure. Think of all the small operations involved in replacing the transmission in your car, and how confusing schematic drawings of such an operation can be. Having access to video of the process being performed, allowing rewinding when necessary, would naturally be much more helpful for a car mechanic uncertain about the details of the process. I have even heard that a telecom company has recorded various repair and maintenance procedures for cell phone base stations on video and com‑ pressed the files to their limit. Should service people find themselves far into the wilderness and facing a repair problem not anticipated, the video clip can

Not referred

146 customers to service desk

Customer goes to service desk Not referred

Simple requests solved

Referred customer

Customer referred

Initial assessment Referred of needs at service desk

15 customers

Phase 2: Initial assessment 26 customers

Figure 3.20  Flowchart with bank customer movement statistics.

Customer comes to bank

Referred

Directly to cash counter

No customer representative available/appointment

Phase 1: Reception

Referral to corporate support

Referral to insurance, property, etc.

External referral of customer

Needs assessed

Assessment of needs in savings department

Needs assessed

Needs assessed

Need for other services

Needs assessed

Needs assessed

Needs assessed

2 customers

3 customers

10 customers

90 customers

Self-service solution used

Assessment of needs in credit department

Referral of customer

Assessment of needs “daily bank”

Needs assessed

Self-service recommended Assessment of needs in cash desk

Directed to selfservice

Phase 3: Sales Customer not relevant for the bank office

Understanding Your Current Business Processes  61

62  Chapter Three

be downloaded to the cell phone and viewed. A hospital has taken a slightly different approach by videotaping all operations. If a surgeon experiences a problem during the surgery, she or he will afterward record voice narrative on top of the video clip to describe the problem, how it was handled, whether the actions taken were successful, and so on. The clips are indexed with descrip‑ tive keywords and stored in a database. When another surgeon prepares for a certain type of operation, all video clips pertaining to it can be retrieved and reviewed. This way, the surgeon will have fresh information about possible problems and how to deal with them.

3.12  Paper and Pencil or PC? A highly relevant and general question for any type of process modeling and con‑ struction of flowcharts is what medium should be used. Is it better to stick to the old tools of paper and pencil or be more up to date and draw on a PC? I believe the answer for this one is both. In the first phase, trying to agree on how the process truly looks, it is definitely an advantage to use paper. In the type of plenary sessions common in this phase, a flip chart or other type of large drawing surface hung on the wall should be used. Usually, many opinions surface, and the flowchart will be a dynamic representa‑ tion that is constantly changing. A technique that enables quite painless adjustment to these dynamics is using colored sticky notes to represent process elements. As the chart is being developed, the notes can be moved around, and the bother of erasing and redrawing is avoided. For this purpose, a computer is of comparatively little use. First of all, the screen is too small. Even if it is enlarged by using an overhead projector, it is still awkward working on a PC in this type of meeting. Furthermore, I’ve learned from experience that there is always a danger that the software tool itself takes attention away from the process and the modeling task. In the next round, however, it is necessary to capture the flowchart from the less manageable, large format used during its development. The flowchart must be stored, copied, and modified. For these purposes, the computer is clearly highly suitable. There are a number of software tools available on the market that make the construction and adjustment of flowcharts very easy. After the flowchart has been completed, but also between meetings during the process modeling, a PC is used to draw the flowchart. The chart can then be printed, copied, and distributed. In addition, the file itself can be sent by e-mail to others who might need the chart (for instance, benchmarking partners). The software allows the user to update the chart as meetings are held. Afterward, the chart can be electronically stored, updated, and changed as the business process

Understanding Your Current Business Processes  63

changes over time. It is difficult and bothersome to do this maintenance by hand, as it involves a lot of erasing and rewriting each time the chart is updated. Such a ­two-­step approach, where the more suitable medium for different pur‑ poses is used, is by far more favorable than categorically selecting one or the other for the entire job.

4 Using Performance Measurement in Business Process Improvement

O

ne argument for the importance of process modeling is that to improve something, you need to know the current state of things. Correspondingly, the general argument for performance measurement is that to improve a process, you must know how well it is performing today. This chapter gives a brief overview of the most important principles of performance measurement.

4.1  The Role of Performance Measurement in Business Process Improvement Trying to run an organization without access to relevant and pertinent information on performance is like trying to drive a car, fly an airplane, or command any other type of vehicle without any instruments. The information provided by the differ‑ ent types of gauges, warning lights, and alarms in a typical instrument panel in a vehicle gives you the decision basis needed: the oil pressure lamp lighting up tells you to check the cause immediately and probably add oil, the speedometer tells you your speed, and a buzzing means you’re driving over rumble strips and you should make sure you’re not veering off the road. Although the performance information you get in an organization is not as up to date and does not require such immediate action, it is still invaluable. An airline with delays on a flight needs to know exactly how many seats are available on later connections for passengers who will miss their originally booked flights. A sales clerk in a shoe store might find that the store has sold out of a pair of boots in a cus‑ tomer’s size, but by checking the inventory of nearby stores in the same chain, he can reserve a pair at a different store for the customer. When a purchaser in a manu‑ facturing company sees that the delivery precision of a supplier has been gradually slipping during the last two months, she knows it’s time to either approach the supplier and demand improvements or look for an alternative supplier.

65

66  Chapter Four

The list of examples is endless, and there are huge differences in how the information is provided, how it is used, and so on. There is no doubt that such performance data provide extremely useful information for deciding to initiate an improvement project. When investigating how performance measurement is used in organizations, we came up with a long list of different application areas (Ander‑ sen and Henriksen, 2004): • Monitoring the overall performance levels of the organization • Setting a strategic direction and using measurement to ensure adherence to this direction • Using average or absolute performance levels to perform detailed opera‑ tional planning of activities and processes • Using performance history to develop cost estimates for products or services • Basing planning on ­up-­to-date performance data • Establishing foresight/early warning through monitoring leading perfor‑ mance indicators • Undertaking regular supplier assessments • Exploiting performance measurement for altering the behavior of indi‑ viduals, groups, or whole organizations and thus used to promote desired changes • Establishing incentives through focusing on certain performance parameters and using performance data as a basis for bonuses or other rewards • Using performance measurement to determine what processes need improvement • Evaluating improvement projects to monitor whether they deliver their promise • Exploiting performance data for marketing purposes • Sharing data with customers, for example, as an approach to documenting that performance is taken seriously • Benchmarking As you can see, most of these have some relevance for improvement work, and we could add to this list the fact that existing performance data about a process can also give useful insight into the causes of inferior performance. This can again be used in an analysis phase in an improvement project to understand which aspects

Using Performance Measurement in Business Process Improvement  67

of the process need changing. In summary, I see a performance measurement sys‑ tem as an integral part of a business process improvement system. By this, I don’t literally mean that you need a clearly defined system for either, but rather that any organization that takes continuous improvement seriously needs to have in place procedures for periodic measurement of important performance aspects. In some cases, this will be an advanced ­ IT-­based system; for other organizations, it can simply mean collecting customer satisfaction data every six months.

4.2  Implementing a Performance Measurement System Although the term performance measurement system has come into general use in management literature, a system can in this context be of any size and com‑ plexity. Thus, this section briefly deals with how to ensure the existence of some kind of systematic performance measurement in your organization to support your improvement efforts. I should, however, mention that during the last few years, more advanced ­computer-­based systems are being used for performance measurement. In my view, many organizations have invested far too much time and money in imple‑ menting systems so complex that they often don’t work, and when they do, there is no real need for all their capabilities. According to Spitzer (2007), 10 percent of a performance measurement system is technology; the rest is how the system is used: the context of measurement, its focus, and the integration and interactivity of measurement. I hope I have made myself clear that this chapter of the book is not a sales pitch for such systems, but rather a reminder that performance information is useful for improvement purposes. A few years ago, a colleague and I wrote a book on designing a performance measurement system. We outlined an ­eight-­step process for designing and imple‑ menting any type of performance measurement system, simple or highly complex. Although I assume you will not decide to design such a system in the middle of a business process improvement project, it might be worthwhile to see the steps such a process can entail. The design process is illustrated in Figure 4.1 (Andersen and Fagerhaug, 2001). The eight steps of the design process are:

1. Understanding and mapping business structures and processes. This is the introductory step of the design process, whose main purpose is to force those setting out to design a performance measurement system to think through and reacquaint themselves with the organization, its competitive position, its environment, and, not in the least, its business processes. You will probably realize that this is similar to what has been dealt with previously in this book: to make sure that improvement efforts are seen in light of the strategy of the organization and its stakeholders.

68  Chapter Four



2. Developing business performance priorities. This step is about under‑ standing stakeholders and aligning the measurements with their needs and expectations. Again, this is identical to what I have talked about as being an introductory exercise to an improvement effort.



3. Understanding the current performance measurement system. If a perfor‑ mance measurement system already exists, then this step is simply a mat‑ ter of reviewing whether parts of this can be reused in the new system.



4. Developing performance indicators. The most important element of your ­state-­of-the-art performance measurement system is the set of performance indicators you will use to measure your organization’s performance and business processes. The purpose of this step is to populate the performance measurement system with an appropriate amount of relevant and precise performance indicators that are able to drive performance.



5. Deciding how to collect the required data. The main purpose of this activ‑ ity is to arrive at solutions for collecting the necessary data for the defined performance indicators. These range from manual data collection to auto‑ matic data capture to picking the required information out of existing IT systems.



6. Designing reporting and performance data presentation formats. In this step, you will decide how the performance data are presented to the users; how they should apply the performance data for management, monitoring, and improvement; who will have access rights to perfor‑ mance data, and so on. To some extent, the design of the presentation formats for the performance data can be compared to the design of the dashboard in a car, where you have to make judgments regarding ergo‑ nomics, the human mind’s abilities for perception and data interpreta‑ tion, and what signals and cues the performance measurement system should provide.



7. Testing and adjusting the performance measurement system. Here, the system is tested so that elements that do not work as planned can be adjusted.



8. Implementing the performance measurement system. In this step, the sys‑ tem is finally put to official use.

It is beyond the scope of this book to deal with this process in detail, so I will leave this process here and instead discuss more specifically the performance indicators that form the heart of such a system.

Using Performance Measurement in Business Process Improvement  69

Step 1

Step 5

Understanding and mapping business structures and processes

Deciding how to collect the required data

1

Step 2

Step 6

Developing business performance priorities

Designing reporting and performance data presentation formats

2

3

Step 3 Understanding the current performance measurement system

Step 7 Testing and adjusting the performance measurement system

?

Step 4 Developing performance indicators

Figure 4.1  The performance measurement system design process.

Step 8 Implementing the performance measurement system

70  Chapter Four

4.3  Performance Indicators To give some insight into the world of performance indicators, it can be useful to outline some typical dipole characteristics. 4.3.1  Hard versus Soft Indicators Hard indicators are pure facts that can be measured directly, whereas soft indi‑ cators are less tangible conditions that must be measured indirectly. Another set of terms is quantitative measures and qualitative indicators. The time it takes to carry out something or how much it costs are typical hard indicators. Quality, expressed as satisfaction of needs or attitudes, is one example of a typical soft indicator. Some differences between hard and soft indicators are listed in Table 4.1. Hard indicators are by far more widely used; soft indicators are seen by many as being so inaccurate that they are rarely useful. On the other hand, as Deming (1986) underlined, the most important numbers are often not known. Management by numbers is one of the deadly diseases that have ruined many enterprises in the Western world. Customer satisfaction is a good example of a soft performance indicator that is best expressed as the customers’ attitudes toward the product or service being delivered. In fear of this being too difficult to measure accurately, many companies have tried to define customer satisfac‑ tion as the number of complaints or warranty costs. To assume that those who do not complain are satisfied is at best naïve and at worst completely wrong. The conclusion is that both hard and soft measures are necessary to give a complete picture. Table 4.1  Differences between hard and soft indicators. Hard indicator

Soft indicator

Objective reference

Observer bias

Accurately known

Surrogate indicator

Hierarchical

Multivariable situation

Example In many cases, when trying to measure softer dimensions of perfor-­ mance, it is necessary to approach the problem through the use of sur-­ rogate indicators. If, for example, there is a need to take a soft measure of quality of the atmosphere in a meeting room, this can be quite dif-­

Using Performance Measurement in Business Process Improvement  71

ficult. It is possible to measure specific conditions such as temperature, humidity, and air circulation. The problem with these measures is that no optimal value can be easily defined, as this will heavily depend on the individual participants in the meeting. A number of surrogate indica-­ tors can be used in this case: • The number of people active in the discussion • The number of ideas or suggestions generated • The number of people leaving the room for various reasons Even if they are not direct translations of what was originally attempted to be measured, such indicators can give indirect interpretations about its performance level.

4.3.2  Financial versus Nonfinancial Indicators Financial indicators include both basic and derived indicators of a financial char‑ acter and/or using monetary values as the measurement unit. Such indicators are almost always hard, but are usually the result of some type of calculation. A few such measures are shown in Table 4.2. Focusing on such financial indicators is obviously an important part of the tra‑ ditional way of running and managing an organization. Financial indicators were often seen as synonymous with performance due to the direct link to the company’s financial result. The increasing understanding that competitive advantage is more closely linked to operational conditions such as quality, service, environmental issues, and so on, has often been met with a redefinition of such operative dimen‑ sions to financially based indicators. Nonfinancial indicators are common denominators for performance indicators that have measurement units other than monetary value. They can be both hard and soft, as shown in the following examples: • Delivery precision • Defect rate • Number of complaints • Customer satisfaction • Quality of work life • Innovation rate • Brand image strength

72  Chapter Four

Table 4.2  Examples of financial indicators. Financial indicator

Calculation (simplified)

Profit margin

Total sales – total costs

Value added

Sales – input goods

Turnover of capital

Sales/total capital

As for hard and soft indicators, the conclusion with regard to financial and nonfi‑ nancial indicators is that both are needed. Problems arise when one of the catego‑ ries is left out of the equation. 4.3.3  Leading versus Lagging Indicators Again, these dipole types of indicators are not necessarily conflicting; rather, they are two types of indicators that should act in a balanced interplay. These two dimen‑ sions are, by the way, often used to illustrate the differences between Western and Japanese thinking when it comes to management. Western management culture emphasizes results and measures accordingly, that is, using measurement systems focused on measuring achievements. In line with traditional Japanese attitudes, the most important part is per‑ forming the process in a sound manner, which will in turn give the desired results. This is reflected in the Japanese measurement systems, where far more emphasis is put on process measures—measures that describe certain important characteristics of a process and that are assumed to have an effect on the desired results. An example is the number of meetings held in a ­cross-­functional team; the typical corresponding Western measure would be the number of changes implemented. Kaplan and Norton (1996) coined the terms leading indicators and lag‑ ging indicators in connection with the balanced scorecard, which is one specific approach to performance measurement. Lagging indicators are aimed at register‑ ing achieved results. By their very nature, they appear quite late in the ­cause-­andeffect chain along a business process, when that process has produced its output (or often has not, which will cause the alarm to sound). The idea is that measuring results is too late; rather, you should measure issues, which will give you an earlier warning that the process is not performing as intended and thus will not produce the intended results down the line. If you listen to certain types of athletes—especially those active in individual, ­concentration-­based sports such as high jumping, speed skating, or bowling— you will hear them talk about something similar. In interviews before important events, they will often say that their strategy is to focus on their tasks and the

Using Performance Measurement in Business Process Improvement  73

Too late! Quality Cost

Flexibility

Env.

Time

Knowledge development

Business processes

Market performance

Financial result

Figure 4.2  Performance measurement as an early warning system.

process, and the results will come. If a bowler is dead set on bowling a strike, chances are high that he or she will fail. But if the bowler focuses on the ­in-­run, the arm movement, and the hook technique and makes sure that these are done well, he or she will not be so tense from the pressure of a strike and the likelihood of getting one will increase. Figure 4.2 illustrates this early warning and attention principle in an orga‑ nization. A simplified ­ cause-­and-effect chain starts with the development of knowledge inside the organization. This knowledge, in turn, is employed to develop and perform business processes, which then create products or ser‑ vices that achieve some level of performance in the marketplace. This leads to a financial result, either good or bad. Typically, there is a considerable time lag from the initial knowledge development to the financial result. The organiza‑ tion will have had plenty of time to make many bad decisions before you can read about them in the financial statement. Therefore, the accounting system is not a good tool for monitoring performance. Focusing on market performance gives you an earlier warning, even earlier if you measure your business process performance. Perhaps the earliest warning comes from monitoring your basic knowledge development inside the organization, and it is often a year or two ahead of when the accounting red light goes off.

74  Chapter Four

Example A manager of a large agricultural collective in the former Soviet Union won the prize for the most productive collective three years in a row. The perfor-­ mance measure used was the number of kilograms of meat produced per year. The fourth year, he shot himself. He had no breeding stock left.

These different dipole types of performance indicators show that performance must be measured in a balanced manner, which is the main message of the bal‑ anced scorecard concept. It also means covering a number of different performance dimensions. If you measure only time consumption in your business processes, there is a danger that employees will sacrifice quality or politeness toward custom‑ ers to achieve this one dimension. Measure only quality and you can end up with a sluggish process that strives for sheer perfection all the time, much more so than what is really necessary. Important performance dimensions include the following: • Time, with regard to speed of delivery, execution, and so on, is becoming more and more important (Stalk and Hout, 1990) • Quality, measured as the more specific factor of defect rate, but also as the product’s or service’s ability to satisfy needs and expectations • Cost, as one important dimension of the performance picture but not the entire picture, as it has traditionally been • Flexibility, which is also a dimension of growing importance; it is expressed as the extent to which tailored solutions can be delivered or volume fluctua‑ tions can be easily handled • Environmental impact, which is becoming crucial when discussing the overall performance of an organization • Safety, for employees and others who come into contact with the organiza‑ tion and its business processes • Business ethics, which over the last few years has proved an extremely important aspect, where failure to uphold ethical standards has brought down several ­well-­known companies The next step is to decide which performance targets are critical to achieve and which improvements this requires.

5 Creating a Business Process Improvement Road Map

B

efore commencing an improvement process, it is important to have a plan. Randomly starting off in an arbitrary direction will rarely lead to good results. This chapter deals with the improvement road map in general and more specific tools that can be used to create it.

5.1  Improvement Project Planning You have perhaps noticed that I have referred to improvement projects many times but have not elaborated on the topic any further. I will not spend much time dis‑ cussing it here, either, but it is worth noting that most improvement efforts are normally carried out as projects. It is possible to envision less extensive improve‑ ment efforts, perhaps involving only one or two persons, not being defined as a project, and in some cases it might be the right approach. However, for slightly more complex undertakings, I definitely recommend organizing and managing the improvement work as a project. This also means resorting to normal project planning practices, that is, defin‑ ing objectives, deriving required activities to attain the objectives, estimating dura‑ tion and cost of activities, and assigning responsibilities. For all these issues, there are countless good project management books available. Thus, this chapter does not deal with traditional project planning or management. Rather, the purpose here is to address how to ensure that the investments you make in improvement efforts are aligned with strategic priorities, with the needs of stakeholders, and with current performance shortcomings. There are several tools and techniques available for this type of analysis. Fortunately, these also

75

76  Chapter Five

c­ omplement one another by addressing different aspects or viewing available data from different angles. The tools that will be covered here are: • Trend analysis • Spider chart • Performance matrix • Criteria testing • Quality function deployment (QFD)

5.2  Trend Analysis Trend analysis, as the name implies, is simply an analysis of how results or perfor‑ mance develops over time. By comparing current results with past performance, it is possible to form an opinion of the direction of the development and sometimes even extrapolate to predict future levels. The latter is rarely very accurate, however, and the main application is to spot trends to enable staying on top of any perfor‑ mance development, be it positive or negative. This is part of the early warning dimension discussed earlier, as an indication of a negative trend usually appears before the performance level reaches an unacceptably low level, thus allowing you to counteract the development with improvements at an early stage. There is really nothing fancy or complex about trend analysis—the main point is to include a sufficient amount of historical data to create a credible trend image. If not enough earlier measurements are included and minor fluctuations are inter‑ preted as a trend, you could make the unfortunate mistake of tampering with a process that is displaying only normal variation. Stepwise, this means:

1. Decide which performance dimensions or indicators to include in the anal‑ ysis and whether to construct one shared diagram for all the data or several individual diagrams. The advantage of including several indicators in the same chart is that it allows identification of any covariance between the indicators; that is, they develop either in rhythm with or in ­anti-­phase of each other. This can be important information, as there might be a ­cause­and-effect relationship between them.



2. When constructing the trend chart(s), time is assigned to the horizontal axis, which places the most recent measurements at the right side of the chart.



3. The performance level is tracked on the vertical axis by plotting the avail‑ able data to form trend lines.

Creating a Business Process Improvement Road Map  77

Delivery time

Performance

Number of new products

Number of bids won

Time

Today

Figure 5.1  Trend analysis for three example measures.

Figure 5.1 shows an example of a simple trend chart, where three measures have been included. Measures that display a negative trend will obviously be relevant candidates for improvement. As will be demonstrated a little later, however, it is not only the development over time that impacts this decision, but also how important the business process is for the competitiveness of the organization.

Example A ­medium-­sized company that strongly emphasized both product devel-­ opment and efficient manufacturing of goods, mainly for export, had been conducting performance measurement for quite some time. The company was not certain that the measurements gave a correct picture of the situa-­ tion. Central performance measures had traditionally been: • Manufacturing costs per unit, both for components and for finished products, measured as the total running costs for machinery and equipment divided by the number of manufactured units. This mea-­ sure had, in previous years, shown a steady increase, and much effort had been put into reducing it. • The costs for purchased parts present in the end product. This measure had also been increasing, and many suppliers had been replaced, without giving the desired results. The company decided to start measuring a number of key performance aspects every six months. Some important performance indicators were: • True manufacturing costs per unit, measured as the manufacturing time multiplied by an hourly rate for the machine or equipment in question

78  Chapter Five

• Price development for purchased parts • The number of new components specified during product development After two years of measurement, a trend analysis of these measures showed a completely different situation, as shown in Figure 5.2. It appeared that the total volume, expressed as the number of products, had decreased somewhat, as one product could now perform more than one function. Thus, the total running costs for the manufacturing equip-­ ment had been distributed over fewer units, which resulted in the believed increase in manufacturing costs. The fact was that regular improvements in this process had decreased the costs. At the same time, the old per-­ formance measure of price development for purchased components had increased because the product developer was now using more purchased parts in the products. In fact, the prices for comparable components had been continually reduced. The reason why the manufacturing costs and the prices for purchased parts had not decreased even more was the con-­ tinual introduction of new components during product development. All these matters were revealed through the analysis, and the improvement effort was directed toward the product development process. Manufacturing costs

Cost/ unit

Price development for purchased parts

Cost/ unit

Number New components during product development 6/05

12/05

6/06

12/06

Figure 5.2  Results of the trend analysis.

5.3  Spider Chart Whereas trend analysis renders it possible to compare the current performance level with previous measurements, the spider chart is an analysis tool offering additional capabilities for graphically displaying your performance data. It is more of a general chart type that has many applications, but in this respect, there are primarily two useful ways to employ a spider chart:

Creating a Business Process Improvement Road Map  79

• To gain a quick overview of the performance levels for a number of different performance indicators simultaneously, mainly to find which are in order and which are lagging • To compare the organization’s own performance level with that of other organizations—a graphical presentation of benchmarking data Figure 5.3 shows an example of a spider chart for the business process of product development. Each spoke in the chart represents one performance indi‑ cator for this process. Obviously, it is also possible to let the spokes repre‑ sent processes and design one chart to represent an overall view of the entire organization. To construct a spider chart, follow this procedure:

1. Collect data from market analyses, surveys, competitor analyses, and so on.



2. Assign one variable to each spoke in the chart.



3. Divide each spoke into logical segments by using a separate unit of mea‑ surement for each variable. The farther from the center of the chart, the higher the performance.

Time to market

Performance level of the most important competitor % new products

Etc.

Cost per new product

Etc.

Market share

Etc.

Etc. Own performance level

Figure 5.3  Example of a spider chart.

80  Chapter Five



4. Plot the performance data for each variable along the correct spokes, using different colors or symbols to separate data points from those of different organizations.



5. Draw lines between the data points for each organization to generate per‑ formance profiles.



6. Identify the variables that show the largest gaps between your organization and the benchmarks.

By plotting the performance level of both your own organization and one or more other organizations, you can see how good your organization is. Depending on where the gap to the competitors is largest, you can select the business processes that should be improved.

Example After realizing where improvements were needed, the company in the pre-­ vious example wanted to compare its status with the status of its com-­ petitors. Through gathering information from different sources, including brochures and statistics from the competitors themselves, a spider chart for comparison was drawn. The result is shown in Figure 5.4. Manufacturing costs

Own performance level

Competitor B

Competitor A

New components specified

Figure 5.4  Spider chart for the central performance measures.

Procurement prices

Creating a Business Process Improvement Road Map  81

This confirmed that the efforts to reduce the manufacturing costs had given results, but there was still a potential with regard to procurement prices, perhaps by entering into more ­long-­term relationships with a few suppliers. Furthermore, the performance level pertaining to standardization of components in new products was dismal. Thus it was correct to start improvement initiatives within this process.

5.4  Performance Matrix The main problem with employing the spider chart for deciding which performance indicators or business processes to improve is that all entries included in the chart are assigned equal weight. From the spider chart, the mechanical approach is simply to conclude that the process needing improvement has the largest gap to the benchmark‑ ing partners. To balance this picture, the performance matrix adds the importance factor to the analysis. This supplements the two preceding tools, which focus on performance alone. Using the performance matrix with the other tools can avoid wasting resources to improve processes that are performing poorly but that are also not very important. A generic performance matrix is shown in Figure 5.5. The matrix is divided into four quadrants, with importance placed along the horizontal axis and the cur‑ rent performance level along the vertical. The performance indicators to be ana‑ lyzed are plotted in the matrix according to the performance data and a subjective evaluation of their importance. The meaning of the quadrants is as follows: • Unimportant (low importance, low performance): The performance level is low, but the low importance renders it unnecessary to put any resources into improvement. • Overkill (low importance, high performance): The performance level is high, but this is of less consequence because the business processes in this quadrant are not especially important to the organization’s competitiveness. Thus, this is not a candidate for improvement. • Must be improved (high importance, low performance): This is the obvious area for starting improvements. The business processes that fall within this area are important, but the current performance level is low. • OK (high importance, high performance): A golden rule is that areas where the performance is already good should also be improved. However, the busi‑ ness processes that in addition to being important are not being performed well today (Must be improved) should be improved first. If no processes fall within that quadrant, processes in the OK quadrant can be relevant candi‑ dates for improvement efforts.

82  Chapter Five

Current performance

9

OK

Overkill 5

Unimportant

1 1

Must be improved 5

9

Importance

Figure 5.5  A generic performance matrix.

The steps in using a performance matrix are the following:

1. Construct an empty chart by placing importance on the horizontal axis and current performance on the vertical axis and dividing both axes into nine segments of equal size.



2. Decide which business processes or other items to include in the analysis.



3. Place each business process in the chart according to its position along the two axes, using symbols to identify each factor.



4. Divide the chart into four quadrants approximately at the middle of each axis. If many items are clustered in one area, place the division lines far‑ ther to one side.



5. Determine which factors fall within the different quadrants.

Example A manufacturer of heavy mechanical equipment for marine applications had identified six critical issues measured through ­ self-­assessment. All mea-­ sures showed a potential for improvement, but there were not sufficient resources to start with all six of them at once. The company found it difficult to decide where to start the improvement activity. To prioritize, a perfor-­ mance matrix was constructed that included the following six measures: 1. Advanced and flexible product design 2. Delivery time

Creating a Business Process Improvement Road Map  83

3. Delivery precision 4. Price 5. Accessibility for maintenance and repairs 6. Product design and finish The matrix in Figure 5.6 helped the company decide to concentrate on improving delivery time and delivery precision.

9

1

Current performance

5

4

6 5

3

2 1 1

5

9

Importance

Figure 5.6  Performance matrix for the six performance measures.

5.5  Criteria Testing Before the criteria testing tool itself is presented, the term critical success factor (CSF) should be defined. One definition is: A limited number of factors that to a large extent impacts the organiza‑ tion’s competitiveness and its performance in the marketplace. Typical examples of such CSFs are the prices, quality, and special features of the organization’s products or services; reputation; and so on. If asking the question, what is it that our customers truly value about our enterprise and helps maintain them as customers? the answer will usually constitute the CSFs. If, however, one tries to improve these directly to increase the competitiveness, it can be difficult to determine where to concentrate the efforts. For example, deliv‑ ery precision might be an enterprise’s main competitive advantage, but the reason

84  Chapter Five

for what causes this and how it can be improved is likely not obvious. Criteria test‑ ing is a ­matrix-­based tool that assists in determining this. A typical form used for criteria testing is shown in Figure 5.7. The procedure for using criteria testing is as follows:

1. Place the identified CSFs, typically three to five, in the upper field of the matrix. If desired, assign each of these a different weight factor that expresses relative importance. In Figure 5.7, weight factors from one to three are used, but other numeric values can be used as well.



2. In the ­left-­hand field of the matrix, place all possible business processes assumed to have an impact on these factors.



3. For each business process, assess its impact on each of the CSFs. Again, the example uses impact factors from one to three, where one is low impact and three is high impact.



4. Multiply the impact factor by the weight factor of the CSF and place the product in the correct matrix cell.



5. For each business process, add up the products and place the total in the ­right-­hand column of the matrix. This numeric value indicates the col‑ lective impact of the business process on the complete set of CSFs. The higher the score, the better reason to improve this process, as this will give the highest overall effect on the organization’s CSFs.

It should also be pointed out that the assessments made during the criteria testing, both to assign weight factors and to determine the impact factors, are quite subjec‑ tive. The numeric values obtained at the end of the test should be viewed only as

CSF

1

2

3

4

5

Weight

3

1

1

3

2

Total score

Process 1

3

1

2

9

4

19

Process 2

9

3

1

3

2

18

Process 3

9

2

3

6

6

26

Process n

3

2

2

3

6

16

Processes

. . . . . . .

Figure 5.7  Matrix for criteria testing.

Creating a Business Process Improvement Road Map  85

directional and not as absolute answers. The more people who participate in the test, the more valid the results are assumed to be.

Example A chain of hair salons initiated an improvement project to rectify its decreas-­ ing sales over a long period. The following CSFs were identified: 1. Price was important, but only as long as it was in line with that of competing salons 2. Good appointment availability, that is, getting an appointment at the desired time 3. The possibility for establishing a ­ customer-­stylist relationship. Because of some recent staff changes, this had been difficult over the past two years, and the customers had reacted negatively 4. Additional services such as a solarium and manicures, as requested by many customers These factors were weighted, and a criteria testing was performed. The result is displayed in Figure 5.8. The conclusion was to concentrate the effort around the three processes that ended up with the highest scores: recruiting, scheduling, and accepting reservations.

CSF

1

2

3

4

Weight

1

3

3

2

Total score

Providing haircuts and other services

3

3

0

0

6

Scheduling

1

9

6

0

16

Accepting reservations

0

9

6

0

15

Recruiting

1

3

9

6

19

Keeping the staff

1

0

9

0

10

Planning additional services

0

0

0

6

6

Purchasing accessories

3

0

0

2

5

Business processes

Figure 5.8  Criteria testing for a hair salon chain.

5.6  Quality Function Deployment The quality function deployment (QFD) tool was developed to represent a ­customer­oriented approach to product development. For this use, it is a methodology for

86  Chapter Five

structuring customer needs, expectations, and requirements and translating these into detailed product and process specifications. The principles can, however, also be used for a number of other problems, including improvement planning. Before showing how QFD can be applied to improvement planning, the theory behind the tool is presented. To start with, the product development process consists of several sequential phases (Akao, 1990): • Transforming customer requirements into a product concept • Transforming the product concept into a product design • Transforming the product design into a process design • Transforming the process design into production documentation Each step of this process must adhere to the original customer requirements. The basic structure of QFD is a relational matrix at different stages of the process, as shown in Figure 5.9. What constitutes the goals of the analysis, which in the first phase of product development are customer requirements and expectations. How expresses the means to reach these goals, which are technical product concepts in the product development’s first phase. In the next phase, these will form the Whats, and How will represent detailed design solutions for the product concept. If a weight factor for each element of What is multiplied by a grade indicating how well each How element satisfies the requirement, an indicator for the performance of each How element is generated, which is placed in the field How much. This way, each phase of the product develop‑ ment process is linked as a chain of relational matrices, as shown in Figure 5.10, to ensure that the voice of the customer is transmitted throughout the entire process. Additional information can be added, thus creating the chart popularly known as the “house of quality,” as shown in Figure 5.11. The QFD process is conducted by entering data into each room of the house of quality. As was mentioned, What represents the external requirements—in prod‑ uct development these are the customer’s product requirements. A weight factor expressing the element’s importance is attached to each element in What, which

How

What

Relational matrix

How much

Figure 5.9  The basic structure of QFD.

Creating a Business Process Improvement Road Map  87

Product concept Customer requirements

Product design Product concept

Process design Product design

Production documents Process design

Figure 5.10  A chain of QFD charts.

What

Importance

How vs. How How

Why

Relational matrix

Points

How much

Figure 5.11  The house of quality.

renders it possible to emphasize some requirements more strongly than others. Why represents the challenges facing the organization—for example, competitors’ solutions to these customer requirements. This can be expressed through a type of benchmarking of different competitors’ offers and their quality or performance. An example is shown in Figure 5.12. After determining how the customer requirements can be fulfilled—that is, How—the relational matrix linking What and How is completed. To make the matrix as clear as possible, it is usually preferable to use as few types of rela‑ tions as possible. A set of commonly used symbols is shown in Figure 5.13. In the same manner, the roof of the house of quality forms a relational matrix to be used for investigating whether there are any relationships among the different ­elements of How. In this matrix, it is possible to indicate both positive and negative relations—that is, factors that work together or that create ­trade-­offs or conflicts. Some common symbols for this matrix are shown in Figure 5.14. For each How, the weight for the determined relationship to the individual element of What is

88  Chapter Five

Poor 1

Good 2

3

4

5

6

7 Our product Company A Company B

Figure 5.12  Benchmarking of others’ offers.

multiplied by the corresponding factor for importance for each requirement ele‑ ment. All products are summarized and placed in the lower field of the chart, How much. Elements of How with a high score in this field should be preferred to others if all elements cannot be incorporated into the product. In improvement terms, QFD is a technique useful for ensuring that the overall requirements and the organization’s strategy are maintained throughout the entire improvement planning process, the same way the voice of the customer is main‑ tained throughout the entire product development process under ordinary use of the tool. For this use, the steps of the process are the following:

Relation

Symbol

Weight

Weak

1

Medium

3

Strong

9

Figure 5.13  Symbols for the relational matrix. Relationship Strong positive Positive Negative Strong negative

Figure 5.14  Symbols for the roof matrix.

Symbol

Creating a Business Process Improvement Road Map  89



1. Identify the requirements to be used in the analysis and place these in the Whats field of the house of quality. The requirements typically stem from strategic decisions, stakeholder expectations, or market surveys.



2. Assign importance factors to each of these requirements.



3. Identify the possible business processes or other elements that are candi‑ dates for improvement projects, and place these in the Hows field.



4. For each How element, assess its impact on each of the What elements and assign an impact factor, normally 1, 3, or 9.



5. Multiply the impact factor by the importance factor for the What element, and write the result in the appropriate place in the relational matrix at the heart of the house of quality.



6. If using the roof of the house of quality, go through each cell in the matrix at the intersection of two What elements. Analyze whether the two ele‑ ments influence each other, and if so, place the appropriate influence sym‑ bol in the cell.



7. If using the benchmarking part of the house of quality, identify relevant data that can be used and establish performance profiles for each data source. Go through each What element and indicate the score of each organization being benchmarked, including your own, on the performance scale. Com‑ plete the profiles by drawing a line along the various data points.



8. For each How element, add up the scores down the column of the rela‑ tional matrix and write the sum at the bottom of the Absolute importance row.



9. Calculate the percentage of the sum of the scores each How element accounts for, and write the number below the scores in the Relative impor‑ tance row.



10. If using the assessment of how difficult it will be to undertake an improve‑ ment project for the different business processes, assign a difficulty factor to each How element in this row.



11. Divide the scores for Absolute importance by the difficulty factor for each How element.



12. Calculate the adjusted Relative importance scores.

90  Chapter Five

Example In a move to reduce the costs incurred for the mother company, PBI, an internal IT department in a large company, was forced to take on external customers. To position itself for this new market situation, the IT department decided to use QFD to identify those business processes it would need to excel at in the future. The resulting QFD chart is shown in Figure 5.15. First, the performance requirements were mapped using stakeholder analysis and through reviewing the redefined strategy for the unit. Each requirement was assigned an importance rating, and a limited competi-­ tive benchmarking was completed from insight into two competitors in the external market. Next, the identified business processes were inserted into the How field. Impact symbols were assigned to each of the requirement/business process combinations of the relationship matrix (empty fields indicate no impact). The products of the importance ratings and impacts were generated and summed up for each business process by using the cor-­ responding numerical impact factor for each of the impact symbols. Just to make sure this is clear, let me explain this in more detail for one of the business processes. For user contact and complaints handling, the cal-­ culation is as follows: • There is an impact factor of 9 for average system ­up-­time, which carries an importance rating of 10; this gives a product value of 90. • An impact factor of 1 for annual cost reduction, multiplied by the importance rating of 5 gives a value of 5. • An impact factor of 9 for external customers’ system ­ up-­time requirements, multiplied by the importance rating of 9 gives a value of 81. • A factor of 3 for annual profit margin, multiplied by the importance rating of 4 gives a value of 12. • An impact factor of 9 for image development, multiplied by the importance rating of 10 gives a value of 90. • The sum expressing how much this business process is able to impact the collective set of performance requirements is 278. • Finally, relative importance figures were calculated in the last field of the diagram, giving an impact rating of 18 percent. The resulting relative importance figures for the entire set of business processes made it much easier for the team to prioritize which busi-­ ness processes to focus on: operations, user contact, and customer communication.

Stakeholder requirements

9

Management/ Reasonable/innovative remuneration and benefits employees Good working environment

Owners

External customers

5

4 10 8 5 7

Annual profit margin of 6% Develop image as reliable and cost-efficient supplier Constantly introduce new technology/solutions Capturing external clients for 25% of the revenues Provide an attractive place of work

Relative importance

5

Vendor respectability

Absolute importance

9

System up-time requirements

7

10

Average system up-time of 97%

Importance

Annual cost reduction of 3% for all major services

PBI internal customers

Performance requirements (Whats)

Figure 5.15  QFD chart for an internal IT department.

Strategic requirements

Business processes or other activities (Hows)

Portfolio planning and service design 141 9%

Development

288 18%

Customer communication and sales 156 10%

Operations 342 22%

User contact and complaints handling 278 18%

Financial management 222 14%

Human resource management 150 10%

1

Own company 2

Company A 3

4

Company B

Competitive benchmarking

5

Creating a Business Process Improvement Road Map  91

92  Chapter Two

After seeing the different tools that aid in prioritizing improvement ideas, you can probably appreciate how they all supplement each other. They differ in focus from trend to competitors’ performance to importance and more comprehensive sets of requirements, and they are quite different in terms of complexity. As a contrast to this chapter, which strongly focused on a selection of tools, the next chapter will deal with “softer” issues in business process improvement.

6 Organizing for Business Process Improvement

A

t this stage in the book, and certainly judging by its title, you might have formed the impression that I believe a good tool will fix any problem in improvement work. Although I believe in the qualities of a good tool, I hope I haven’t gone quite that far yet. There are a number of traits that set an orga‑ nization adept at continuous improvement apart from a merely adequate one. This chapter is a “softer” one, focusing on some of these traits: improvement skills and training in them, organizing for improvement, and stimulating an ­improvement­oriented organizational culture.

6.1  Business Process Improvement Skills I’d like to revisit the brief discussion at the start of the book about the apparent contradiction between continuous improvement and breakthroughs. While I often find the frequent sports analogies in management literature to be misplaced, I can‑ not help but introduce one here. Think of improvement capability as a muscle. As you know, a muscle must be regularly exercised to stay fit and be able to perform. Athletes know this very well, which is why they keep exercising, paying special attention to those muscle groups that are critical in their type of sport. It would be highly unlikely to hear any world or Olympic champions assign their victories to three months of doing nothing but eating and smoking prior to the event. The same argument can be made for the “improvement muscle.” If it is left dor‑ mant for long periods of time, it will be difficult to revive every time it’s needed. For example, think of health clubs just after New Year’s. They are crowded with overweight people with resolutions to get fit. You can almost hear their muscles cranking back into life. If any of these individuals manage to stay the course, it will take months before their muscles are functioning smoothly. Quite the opposite of

93

94  Chapter Six

a fit athlete with toned muscles and strong general form. Thus, when starting an improvement project, make sure your improvement skills are ready to go. To achieve this, some preconditions must be fulfilled: • A sufficiently large portion of the organization must be trained in business process improvement and involved in improvement projects. This is not a skill reserved for top- or ­middle-­level management. The best improvement projects I have observed always either include team members from all levels of the organization or are run exclusively by employees at lower levels in the organizational hierarchy. • There must be a sufficient amount of improvement projects going on at any given time so that employees are regularly involved in improvement work. If three years pass from the time a person is part of an improvement project, she or he will have to start almost from scratch every time. This is both a mindset and a utilization of skills that takes time to get comfortable with, so never falling out of it is the preferred option. • The organizational structure must allow employees at all levels the freedom to get involved in improvement efforts. This means freeing up time for par‑ ticipation in specific improvement projects and establishing arenas where people can meet to develop ideas and share experiences. • There must be incentives for engaging in improvement work. The organiza‑ tion depends on individuals or groups of people who initiate new improvement projects, have enough passion for the ideas to see them through to approval, and execute them properly. Usually, this comes in addition to regular respon‑ sibilities in the organization, and unless there are rewards for this extra effort, people will not do it or will be quickly weaned from this type of behavior. The latter two of these will be dealt with in later sections. For the first two, it’s really a question of whether you want to have a constantly fit team of improve‑ ment players available for the next match, or you are content with having to train a bunch of couch potatoes each time the need arises. In assuming the latter is not too appealing, the only feasible mechanisms at hand are training and practice. Training can take many forms and shapes. Traditional courses can be engag‑ ing and powerful, or they can be boring and easily forgotten. Training can take place through ­experience-­sharing sessions among people who have been involved in improvement projects. To allow people more freedom about when and where to acquire knowledge, various e-learning approaches can be used. There have been experiments to develop applications that resemble computer games and that are used for transferring more tacit skills, and whose engagement factor makes them very attractive. Irrespective of which training approaches are used, I would ide‑ ally like to see every single employee receive at least a modicum of insight into improvement work.

Organizing for Business Process Improvement  95

6.2  Organizational Modes That Support Business Process Improvement Because this book is not about organizational design, I will not outline the differ‑ ent philosophies for constructing organizations and what these entail. Rather, I will address a few issues that have a direct bearing on the opportunities and climate for business process improvement. At issue is to what extent the business processes of an organization are visible and appear in the organizational structure. We have already seen how traditional departmental silos hamper flow and cooperation along business processes, but I did not really offer an alternative. There is, in fact, an alternative—namely, struc‑ turing the organization along its business processes. This can take many different shapes, but it usually involves some type of process ownership in the organization. A common problem in organizations still focusing on departments or just starting to orient themselves toward business processes is that very few people feel responsible for the organization’s processes. There are usually limited definitions of where one process ends and the next one starts; no one has been given the responsibility of delivering the product or service from the process to the customer, on time and of the right quality. At best, someone assumes responsibility for the process, but then often more than one person does so simultaneously. The result is that local “emperors” can be found around the organization, ruling over segments of the process, but without having anyone fully responsible for the entire process. A distinguishing feature of processes without ownership is that no improve‑ ments are initiated unless there exists a clear order from management to do so. Process ownership means appointing people as owners of the individual busi‑ ness processes. Management must do the appointing, and a general rule is that the person who has the highest organizational authority in the process or who impacts the largest portion of the process should be its process owner. A process owner is given a variety of responsibilities: • To initiate changes to the process. • To continually measure the performance level of the process. • To initiate improvements in the process. • To establish some type of steering group for the process that will work together to improve it. This steering group will normally consist of people who perform tasks within the process, but suppliers and customers of the process are also relevant members. After a company introduces process orientation and process ownership, the organi‑ zational chart will change character, from vertical silos with department managers to horizontal processes with process owners. Examples of the two extremes are shown in Figures 6.1 and 6.2.

Department co-worker #1

Department co-worker #2

Department co-worker #3

Department co-worker #1

Department co-worker #2

Department co-worker #3

Department co-worker #3

Department co-worker #2

Department co-worker #1

Logistics manager

Figure 6.1  Traditional organization with vertical silos.

Sales manager

Marketing manager

Department co-worker #3

Department co-worker #2

Department co-worker #1

Production manager

CEO

Department co-worker #3

Department co-worker #2

Department co-worker #1

Procurement manager

Department co-worker #3

Department co-worker #2

Department co-worker #1

Financial manager

Department co-worker #3

Department co-worker #2

Department co-worker #1

Personnel manager

96  Chapter Six

Organizing for Business Process Improvement  97

CEO Process owner

Manufacturing and delivery of products to customers Process co-worker

Process co-worker

Process co-worker

Process owner

Process owner

Order handling

Manufacturing

Process co-worker

Process co-worker Process owner

Delivery and installation

Process co-worker

Process co-worker

Process owner

Process owner

Obtaining customer commitment

Invoicing

Process co-worker Process co-worker

Process owner

Product development Process co-worker

Process co-worker

Process owner

Personnel development

Process owner

After-sales service

Process co-worker

Process co-worker

Figure 6.2  Organizing along business processes with process owners.

It should be pointed out that there will often be a hierarchy of business processes and corresponding process owners. As can be seen in Figure 6.2, the extensive process of manufacturing and delivery of products to customers has one process owner coordinating ­high-­level issues related to this process. At the same time, this process consists of several shorter processes at the level below, for example, order handling, manufacturing, and delivery and installation. The individual process owners for these subprocesses are responsible for their isolated processes, but they must coordinate their activity with the process owner of the overall process. I have seen examples of organizations that have taken this model to the extreme, that is, disbanding all traditional departments and assigning all employees to one or more process departments. Process departments are organizational units that have been assembled not according to the principle of grouping people of similar expertise, but rather to contain all the different types of expertise required to per‑ form the process in question. These units are headed by the process owner, who has the human resources management responsibility toward the team members, and sometimes they are shared by two or more process owners in cases where members belong to more than one process team. Although this organizational model carries inherent advantages, there are dis‑ advantages as well. Among these is a tendency not to have room for ­full-­time posi‑ tions in many processes, thus ending up sharing resources between processes. The

98  Chapter Six

line of authority toward these employees is often blurred; in many cases process owners fight over employee capacity in situations of high workload. Furthermore, some process team members claim that they miss colleagues with the same back‑ ground as themselves with whom they can discuss professional problems. Quite a few organizations have tried to strike a balance between the two extremes by designing a type of matrix organization. These still maintain the tra‑ ditional functional departments, where the department managers hold the human resources management responsibility over the department employees. At the same time, the business processes have been clarified, often even drawn into the orga‑ nizational chart, overlaying the departments. All employees in the departments have further been defined to belong to one or more business processes, giving the process owners authority to decide on tasks and how their time is spent. From what I have seen and heard, this does not solve all problems; there are still cases of dif‑ ferent managers pulling employees in their own direction, but this model can be a reasonable compromise. These different models dictate primarily how employees are organized in daily operations. Here, we are more concerned with how they function during improve‑ ment efforts. And if there is one lesson I’ve learned after working with many dif‑ ferent organizations over the years, it is that improvement work is a team sport. Teams bring much more knowledge, ideas, and creativity to the table than does an individual, and there is nothing as frustrating as struggling with an improvement project alone. Frankly, there are no reasons why one single person should carry out an improvement project. There is no secret to composing improvement teams; it is a matter of finding the right people and giving them the necessary mandate for their project. Regard‑ ing members, the team should fill the following roles: • Team leader, that is, someone responsible for calling meetings, monitoring the progress of the project, and making sure that minutes are prepared and that necessary information is gathered and distributed. Many of these tasks can usually be delegated, but the responsibility still remains with the team leader. It is quite common that this role is given to the person in the team with the highest position in the organizational hierarchy. • Link to management, either through management representation in the team or by giving someone authority to represent the management. • Process owner or the person who in other ways is responsible for the whole or most of the process to be improved. • Other people involved in the process, preferably from a ­cross-­section of the organizational units the process runs through. It is extremely important to make sure that the team does not consist solely of people located above or on the side of the process to be improved. If those who will later have to

Organizing for Business Process Improvement  99

change their routines are not involved in the team, making the necessary changes might be hard to do. • An internal or external customer of the process. Along with the process view comes a commitment to viewing the customer of the process as the most important part of the puzzle. By involving the customer in the team, it is possible to make sure that the customer’s views are adhered to in the improvement activity. • An internal or external supplier to the process. In the same way that any process has a customer, any process also has a supplier that provides some sort of input. The supplier is to some extent a part of the factors that shape the performance level of the process through the input being supplied, and should therefore be included in the improvement activity. • Possibly some kind of external assistance in cases where the improvement tool to be used is new to the team. It is possible for one person to take on several roles. From experience, a team should not have more than six or seven members, but this depends on the scope and complexity of the project. At the same time, care should be taken so as not to establish a team that is too small, as this might hamper its ability to perform its tasks, and it could undermine a broader sense of ownership of the project’s results. The following requirements for candidates should be posed when selecting team members: • Time to participate actively and wholeheartedly in the work. Instead of including key personnel who never have time to participate, you should choose people who will make an effort in the project. On the other hand, there are usually some people you cannot do without, and these individuals will have to set aside the necessary time. • Competencies and knowledge about both the organization and the relevant business processes, as well as training in the tool expected to be used. Alter‑ natively, training in the use of the tool must be provided at the outset of the project. • Motivation to work on improving the selected process. A person selected against his or her will is an unsuitable member of an improvement team. • The ability to cooperate, listen, and communicate, as this is a typical team effort that is less suited for the traditional silent loner. • Credibility and respect in the organization, to ensure an impact when pre‑ senting results from the project and proceeding with effective implementa‑ tion of improvements.

100  Chapter Six

Finally, a warning must be issued against recruiting members on the basis of the think‑ ing, “You have nothing better to do; go participate in this improvement project.”

Example A manufacturing company with approximately 400 employees started an improvement project connected to the procurement process and its link to the more general logistics processes. An improvement team for the project was established, and it consisted of the following people: • Procurement manager • Logistics manager • Production planning manager In addition, external individuals were hired to be part of the team. This composition had several negative consequences: • These three managers traveled quite extensively and were generally rather busy, which made it difficult to arrange meetings or to have any work done between meetings. • The fact that they were quite high up in the organizational hierarchy meant that they were less involved in the operative execution of pro-­ cesses. Thus, the knowledge about the processes within the team was too low to enable any detailed documentation of the process or generation of improvement suggestions. • Each member of the team had a large degree of responsibility within her or his own area and felt personally responsible for the perfor-­ mance level within it. It was therefore difficult to admit weaknesses and problems, and very few constructive improvement suggestions achieved broad support from the team. It would have been wise to allow more operative personnel into the team.

In addition to these general improvement team guidelines, there are some more permanent groups whose tasks are to continually work with improvement initia‑ tives. More specifically, two such groups are described here: quality circles and ­cross-­functional teams.

6.3  Quality Circles Quality circles originated in Japan and were highly popular when the phenom‑ enon was first described in Western management literature in the 1970s and

Organizing for Business Process Improvement  101

1980s. They might still be used in Japanese organizations, but I have not been able to uncover any statistics or documentation about this. Many academics and consultants tried to import quality circles to the United States and Europe, and some companies, after finding quality circles to be highly worthwhile, still use them. A quality circle can be said to consist of (Aune, 1985: 11): • A group of people from the same work area (they have the same type of job and experience the same problems) who; • Under the guidance of a circle leader (in the beginning preferably a foreman or the like); • Voluntarily participate in regular meetings during normal working hours, approximately one hour per week; • To, according to their own priorities, identify, analyze, and solve problems within their own work area, and; • In writing and orally present suggestions for solutions with cost estimates to a person authorized to decide on implementation. The objectives of the work in quality circles are, on purpose, defined to be twofold: • To strengthen the competitiveness of the organization through creating improvements • To create a good basis for development of the individual employee’s cre‑ ative skills and allow the use of these skills in practical improvement work The standard approach is to establish several quality circles, each having its own circle leader. Above these is usually a facilitator who can assist if some of the cir‑ cles get stuck and who also functions as a point of contact for the management of the organization. Quite often, we can also find a steering committee for the circles that coordinates the entire activity in this area. The committee usually consists of members from management. When these different elements are put together, the organization of the quality circle work can look like Figure 6.3. I mentioned that a number of organizations in the United States and Europe tried quality circles during the 1970s and 1980s. Studies have shown that there are a number of reasons why quality circles have not worked too well in the West: • The circles were not part of a larger quality program • The circles never became an integrated part of the work in the organization • The circles were introduced before the general consciousness level about quality had become sufficiently high

102  Chapter Six

Circ

le

Circ

le

der

ad

er

Leader

Lea der

Circle

tor Fac i

lita

Le

tor ilita Fac

der

le

Circ

Lea

Steering committee Facilitator

Le

Ci

der

ad

er

Lea

Ci rcl e

Lea

cle Cir

der

Lea

er

ad

Le

e

rcl

rc

Ci

le

Circle

Figure 6.3  Organization of quality circles.

• The companies were more concerned with the financial gains achievable than the working mode itself • The work in the circles was perceived to be too bureaucratic, and all the imposed rules seemed artificial • As the circles succeeded in solving the smaller problems locally, they ran out of problems they could attack The last explanation is especially relevant in terms of how improvement work is organized. There is often an evolution in organizations that take continuous improvement seriously, from first addressing local problems within organizational units to escalating to more complex problems. These are often attacked using ­cross-­functional teams.

6.4 ­ Cross-­Functional Teams Such complex problems are rarely so unidimensional that they can be solved within one organizational unit. In mature organizations that have worked with improve‑ ment for a while, it is not unusual that, at some stage, the remaining problems are of a ­cross-­functional nature, thus requiring ­cross-­functional improvement efforts. True improvement of such problems often results from efforts in the “white” zones

Organizing for Business Process Improvement  103

of the organization chart, that is, at the interfaces between departments or busi‑ ness processes. The answer to this challenge can often be ­cross-­functional teams that possess broader competencies and, not in the least, organizational positions to work on the case. A ­cross-­functional team can be defined as follows: A group consisting of members from different functional departments or areas of responsibility and often also from different hierarchical levels. The purpose of the team is to solve problems involving several organiza‑ tional units or business processes. Cross-functional teams can be formed on an ad hoc basis, where a team is formed to solve or investigate a specific problem and is thereafter dissolved, or they can be permanent, like the quality circles. The normal and recommended size of such teams is between four and eight people. There are also rules that govern the work of ­cross-­functional teams: • Meetings are held only when there is 100 percent attendance • Meetings are held in “sacred” places, that is, not on the home turf of any of the members and somewhere where there are no interruptions • Meetings are held during regular working hours • Before the team is established, management must agree to follow up any rec‑ ommendations made by the team, subject to realistic budgetary constraints • Experts can be summoned according to needs It should be expected that during the first meeting after the establishment of the team, the work will be rather theoretical and characterized by the fact that the team and its members are trying to find their places. Such discussions can seem a waste of time and not leading to any solutions or improvements, but they help educate the members and allow them to get to know each other. For each problem attacked by the team, the result should be a list of improvement actions with related cost estimates, a schedule for implementation, and indications of expected results.

6.5  Stimulating an Improvement Culture When having the privilege of seeing different organizations from the inside and observing how they work with improvements, I find that the single most defining characteristic in truly ­world-­class improvers is the way the organizational culture supports continuous improvement. There is almost a tangible quality to the atmo‑ sphere in these organizations that spells optimism, positivism, and a ­never-­ending

104  Chapter Six

quest for something better. I should also emphasize that this climate does in no way seem desperate or frantic; these organizations have the ­ self-­confidence to know they are already good, but they genuinely believe they can do even better. It is a culture that instills confidence in those who come in contact with these organiza‑ tions, and more often than not inspires them into trying to copy it. Of course, this is not easy. This type of culture cannot be implemented the following week; it is carefully developed and nurtured over a long time (if anything, the only thing you can do quickly is ruin such a culture by starting to behave inconsistently). Studies have reported finding this type of organizational culture in highly suc‑ cessful organizations. The study Good to Great (Collins, 2001) identified a num‑ ber of companies that for a period of at least 50 years had been highly successful, along with a control group of companies in the same industries that had performed only averagely. The study team found that the successful companies shared some similarities (which were found to contribute to either stimulating progress or pre‑ serving the core of the companies): • BHAGs (big, hairy, audacious goals; stimulates progress) • Cultlike culture (preserves the core) • Try a lot of stuff and keep what works (stimulates progress) • Homegrown management (preserves the core) • Good enough never is (stimulates progress) The latter of these capture the spirit I described earlier of ­world-­class ­improvement­oriented organizations and the relentless pursuit of ­ever-­higher performance levels. So how do you create and maintain such a culture? This is a question impossible to answer. There are probably a number of ways you can arrive there, all of which have no quick fix. I cannot offer you a list of steps, either, but I hope the following items can at least play an important role for organizations trying to stimulate such an ­improvement-­oriented organizational culture: • Encourage employees at all levels in the organization. When things are working fairly well, with decent results being produced, it is tempting to rest on your laurels. In such times, employees who launch criticism, point out things that are not perfect, or simply suggest possible improvements can be seen as killjoys. Realizing that following up such ideas will require time, money, and energy, it is easy to kill the idea by dismissing it or asking that it be postponed. This is a ­fail-­safe way of killing any improvement spirit that might have developed. Always be positive toward such ideas and treat them with the respect they deserve. In the end, it might very well be that the idea was a bad one or that pursuing it was not cost effective. But what’s impor‑ tant is that you have shown that ideas are welcome and taken seriously.

Organizing for Business Process Improvement  105

• Celebrate victories. When improvement projects are completed, you should mark the occasion. Those pioneering the project should be “shown off” to the rest of the organization, and the results achieved should be announced. The positive energy created through such success stories is invaluable. • Never allow the organization to become overly content or complacent. In celebrating the victories, make sure that the organization is not lulled into a false sense of security and contentment. There is no organizational culture that produces so few improvements as a complacent one, perhaps with the exception of one that is complacent without any reason for being so. • Expose the organization to continuous benchmarking. There are fewer ways that are more effective in combating organizational complacency than benchmarking. Comparing performance levels of your organization with relevant benchmarks can shake any overly ­self-­confident organization out of this unfortunate state. If you are hovering around three weeks average delivery time for a certain type of product and you demonstrate to the orga‑ nization that a competitor consistently delivers in one week, chances are good that improvement ideas will surface shortly (unless you allow people to argue that the benchmark is not suitable or to give you numerous excuses why you could never achieve similar results). • Reward improvement achievements. Organizational rewards come in many forms: honor and visibility, various perks, money (bonuses or pay raises), and promotions. All these can help motivate people to improvement efforts, including those who have already achieved results and are rewarded for this, and those who see that creating improvements pays off. If you have a perfor‑ mance measurement system in your organization, data from this system can often be useful as input into reward systems. A word of warning is warranted, though. Researchers argue about the effects of different types of rewards, some even claiming they achieve very little. Unfairly awarded rewards might have negative effects on motivation, and awkwardly designed reward sys‑ tems can encourage undesired behavior throughout the organization. This is especially true for individual rewards, which can stimulate an egocentric or selfish culture that hampers teamwork and business process thinking. At this point in the book, I am happy to announce that I have presented all the introductory material preceding the actual business process improvement toolbox. We will now move on to the more detailed improvement tools that you can use to create these coveted improvements.

7 The Business Process Improvement Toolbox

A

long with the business process improvement toolbox, this chapter includes an overview of the different phases in improvement work and the tools that belong to each phase. A far more detailed treatment of each tool is given in the ensuing chapters.

7.1  The Need for a Toolbox It is inherent in the nature of tools that their purpose is to offer support. For exam‑ ple, if you want to fix your car, you select and use the tools that will get the job done. The desired result is a car that runs, and it doesn’t matter whether a wrench, a pliers, or a screwdriver was used. In the same way, the desired results for organi‑ zations are improved performance and competitiveness. For this purpose, they can use whatever tools work, as long as they give results. It is thus far more important to know how to select the right tool and use it to reach the set goals than to know a large number of different tools. At conferences, seminars, and other forums, claims have been made regarding the use of a particular tool. These claims are often followed by an intense debate. Such statements and allegations show that too much emphasis has been placed on the tool itself and not on its purpose. The fact is that all the tools presented in this book pose different requirements to the user, they serve different purposes, their use and results depend on the situation in which they are applied, and they give different results. Thus, it is not possible to allege that any one tool is best. All these tools should be seen as belonging to a large and ­well-­equipped tool‑ box available to the organization striving to improve. They should supplement each other and function in symbiosis. Which tool is used in a specific situation depends on the characteristics of the organization using it and the situation in which it is applied. First, the situation and problem at hand must be defined, and then a suitable

107

108  Chapter Seven

tool selected for it. If we have only one tool—say, a hammer—it is amazing how quickly all problems come to look like nails. The purpose of this book, therefore, is to present an extensive toolbox for the business process improvement handyman.

7.2  A Business Process Improvement Model I presented a framework in Figure 1.2 where I tried to illustrate the different fac‑ tors that are required for or influence business process improvement. We have now dealt with all the items in the framework except for the core activity of improving business processes. To make it more manageable, it is useful to divide the larger task into a number of phases that improvement projects typically go through. This can be done in many different ways, and I assure you it is not entirely easy to cre‑ ate a logical flow of activities in such a model. I have tried to reorganize the stage model presented in the first edition of the book, ending up with the one presented in Figure 7.1. One of the problems in designing such a model, which by nature must be por‑ trayed as linear, is that such processes are rarely so neat. Improvement teams will find themselves skipping phases, retracing their steps, reiterating work, and so on. I hope you will see these stages merely as illustrations of some distinct groups of tasks or mindsets during such projects. The seven phases of this model have the following objectives:

1. Develop improvement priorities based on an overall performance under‑ standing. This is a preliminary stage that precedes specific improvement projects, and the aim is to understand which business processes will be important to improve, given the organization’s situation, stakeholders, and strategy.



2. Understand the current business process and the performance shortcom‑ ing in question. This is normally the first step in an improvement project, where the objective is to understand how well the process works at the moment and what seems to characterize the problem experienced.



3. Collect data about the performance shortcoming. Business process improvement work is at its best when it is based on facts and data, and the purpose here is to gather relevant information for the ensuing analyses.



4. Analyze the performance shortcoming. In this stage you apply different techniques to try to understand the true nature of the problem causing the ­less-­than-desired performance.



5. Generate ideas about causes of the performance shortcoming and possible improvements. This is a creative phase to generate a broad range of ideas

The Business Process Improvement Toolbox  109

7. Implement the improvements 6. Develop improvements to remedy the performance shortcoming 5. Generate ideas about causes of the performance shortcoming and possible improvements 4. Analyze the performance shortcoming

3. Collect data about the performance shortcoming 2. Understand the current business process and the performance shortcoming in question 1. Develop improvement priorities based on an overall performance understanding

Figure 7.1  Stages in business process improvement work.

about what causes the performance shortcoming, but it is also a selec‑ tion of tools that can be used in other phases of the improvement process. In Chapter 10, I have included some techniques for choosing among the resulting ideas and creating consensus in the improvement team.

6. Develop improvements to remedy the performance shortcoming. To many, this is the exciting stage of the process, where the objective is to create new solutions or process designs that will eliminate the performance shortcoming.



7. Implement the improvements. While perhaps sounding easy compared with finding ways to improve the process, implementing lasting change can often be the hardest part of an improvement project.

Histogram

Bottleneck analysis

Is–is not analysis

Flowchart with statistics

Spider chart

Scatter chart

Five whys analysis

QFD

Several-leveled flowchart

Trend analysis

Problem concentration diagram

Cause-andeffect chart

Matrix diagram

Flowchart divided into process segments

Strategy map

Check sheet

Pareto chart

Criteria testing

Crossfunctional flowchart

Competitive forces analysis

Surveying

Critical incident

Analyze the performance shortcoming

Relations diagram

Traditional flowchart

SWOT analysis

Sampling

Collect data about the performance shortcoming

Performance matrix

Relationship mapping

Understand current process and performance shortcoming

Stakeholder analysis with Kano model

Develop performance priorities

Table 7.1  Tools for different phases in the improvement work.

Paired comparisons

Six thinking hats

Nominal group technique

Crawford slip method

Brainwriting

Brainstorming

Generate ideas and choose among them

Benchmarking

Business process reengineering

Six Sigma

Statistical process control

QFD

Idealizing

Streamlining

Develop improvements

Force field analysis

Tree diagram and process decision program chart

AΔT analysis

Implement improvements

110  Chapter Seven

The Business Process Improvement Toolbox  111

7.3  The Tools in the Toolbox Within each of these phases are a number of tools and techniques that I have found to be useful and possible to apply without a PhD. This selection of tools can certainly be discussed. Some people will find that there are tools missing that they think should be there, and others will probably try a tool and find that they don’t like it. I accept this, and I am certainly open to feedback and suggestions for changes. The world is, fortunately, not sufficiently square that all the tools can be singu‑ larly defined for each criterion and thus placed in the right category. The classifica‑ tion has been carried out according to the most prominent features of the tools and in complete confidence that exceptions exist. The presentation of the tools throughout the book will follow the order of the stages, starting with stage 3, as the first two have already been dealt with. For each stage, Table 7.1 lists the tools and techniques included. Depending on which phase the project is currently in and which task you want to solve, this table can serve as a guideline for what tools are available. For further advice on which tool will work in a given situation, I can offer no additional help other than to experiment. There is no doubt that some of these are more complex, require more skills to apply, and take longer to complete; additionally, the degree of change or innovation they con‑ tribute to also varies. However, it seems impossible to draw up general rules about these issues. I suggest trying different ones and seeing which work for you.

8 Tools for Collecting Data about the Performance Shortcoming

A

t this stage of improving a business process, you will typically possess a process model of some sort for the business process you have undertaken to improve. This process has been selected because it displays a perfor‑ mance shortcoming of some sort. You might even think you know fairly well what the problem is and how it can be fixed. If so, a word of caution is in order: ­spur­of-the-moment inspiration is rarely the best starting point for your improvement project. If there is one thing most authorities in the area of quality management agree on, it is the importance of ­fact-­based decision making—knowing the situ‑ ation at hand well enough to make sure you proceed in the best direction. In this chapter, you will learn about different tools and techniques for collecting data and information about the performance shortcoming you are facing. There are four key approaches and tools in this group: • Sampling • Surveying • Check sheet • Problem concentration diagram

8.1  Sampling Sampling is not a specific tool per se, but rather a useful principle to know about before collecting data. While it might sound trivial, data collection can be a ­time-­consuming and expensive task. Sampling is simply a way of economizing the ­data-­collection pro‑ cess. The purpose is to allow collection of a smaller amount of data while maintaining a data set with sufficient validity to draw conclusions about the situation being studied.

113

114  Chapter Eight

There are many ways to determine how the sample should be collected. Some of the most common types include the following: • Random sampling: In this method, random numbers determine which data to draw from the full population. You can find such random numbers in published tables, by using a computer to generate them, or simply by throw‑ ing dice. An example is pulling out customers 4, 11, 19, 21, 34, and so on, among the first 1000 to be served during the month and asking his or her opinion. • Systematic sampling: In many cases, it is difficult or even impossible to pick out random samples. In systematic sampling, data are collected at fixed intervals of time, numbers, length, and so on. For example, every 20 min‑ utes the number of customers waiting in line is counted. • Stratified sampling: For some business processes, there are differences between categories within the entire population. In this case, you should purposely collect data from each of the categories so that the samples rep‑ resent the categories in correct proportion to one another. If a company has seven salespeople, customer satisfaction scores can be collected from the customers of each, relative to the number of customers each serves. • Cluster sampling: If you know that the population is stable and without much variation over time, cluster sampling is the simplest way of defining the sample. In this case, a group of the units is taken to represent the whole population—for example, the entire batch of parts produced in one hour to represent the entire week’s production. As I’ve already mentioned, sampling is not one unified tool, which means it is dif‑ ficult to define generic steps. Rather, sampling is used to support other tools, but there are some important issues to keep in mind:

1. Assess the nature of the population to be sampled to decide on a suitable sampling approach (take into account the homogeneity of the population, any clustering of data, and so on).



2. Decide on the sample size, which can be a tricky task. For help with this issue, consult available standards that specify sample size and frequency. Such standards are available as military standards, ANSI standards, and ISO standards. For a list of relevant standards, see http://www.variation. com/techlib/standard.html.



3. Collect the sample of data according to the chosen sampling approach.



4. Calculate simple figures such as averages and means so you can test whether the sample is a reasonable representation of the population.

Tools for Collecting Data about the Performance Shortcoming  115

Example A large hospital had experienced several outbreaks of infections, and these had occurred in many different units across the hospital. A doctor looking into the problem found that the presence of a certain type of bacterium seemed to predict such outbreaks quite accurately, and that this warning came in time to take measures to avoid or at least reduce the extent of the outbreak. It was therefore decided to start testing for this bacterium, with the hope of being able to stop many outbreaks before they could take hold. In the beginning, the idea was to test extensively in every unit, but it soon became clear that the costs and disruptions were almost a worse problem than the consequences of the occasional infection outbreak. A statistician was consulted and asked to design a sampling plan that would be less of a burden on the personnel and patients, but at the same time would be effective in warning of future infections. She came up with a cycle plan in which each unit would do tests once every second week. After this procedure was implemented, the test proved capable of pre-­ dicting several outbreaks, and many of them were thwarted altogether by speedy action.

8.2  Surveying Sampling gives you advice about how you can collect less data and still be reason‑ ably sure you capture the full picture of the performance shortcoming. In many cases, the data you need to collect are precise and specific—for example, the dimensions of a physical product, the color of a chemical test on a color scale, or the costs of different aspects. However, in many business processes, information about people’s attitudes, feelings, or opinions is more important. Such information cannot be collected using tools suitable for physical dimensions or pH values. For these matters, surveying is used to collect data. Before looking at the steps for conducting a survey, it should be mentioned that there are several different ways of conducting surveys. First of all, the basis for a survey is normally a questionnaire of some sort. The questionnaire outlines the questions to be asked of the respondent, and depending on how the answers are collected, the questionnaire can come in many different forms. There are at least two different ways of obtaining answers from the respon‑ dents. First, the questions can be presented in writing and the answers collected the same way. For this approach, you can use a questionnaire that is mailed or handed out personally, or you can use an electronic one, sent by e-mail or published on a Web site. Second, the questions can be orally transmitted, and the respondents can answer back orally. This can take place in traditional interviews, either ­face-­to-face

116  Chapter Eight

or over the phone, or managed by a computer that reads questions to respondents over the phone and records their answers. In all types of surveys, large amounts of data can be collected relatively easily and inexpensively. While questionnaires filled in by the respondents generate the most data per dollar, interviews tend to give data of higher quality. To conduct a survey, follow these steps:

1. Clearly define the objective of the survey and how the data will be used later



2. Determine what information is required to achieve this objective



3. Decide how the survey will be undertaken—that is, written (via mail, fax, e-mail, or the Internet) or verbal (by telephone or in person)



4. Develop the questionnaire, keeping in mind issues such as type and sequence of questions, understandability, language, grouping of questions, brevity, and so on



5. Test the questionnaire to ensure that all questions are easy to understand and can measure what they are intended to



6. Identify the sample of respondents



7. Perform the survey according to the chosen approach

Example A private retirement home had seen the number of vacant rooms steadily increase over the last few months, and informal talks with both residents and their next of kin indicated that many were dissatisfied with the services offered. However, these brief talks, often just when passing each other in the hallway, gave little specific insight into the cause. The management decided to undertake a more thorough survey to gain a better understanding of how it could turn the trend. It was decided that it was necessary to survey both the residents and their relatives, especially since the latter group often acted as legal guardians of the residents and made the decisions about moving them. Two questionnaires were developed, one for the residents and one for the relatives. Both were short, two pages, and asked the respondents to rate factors like the comfort of the rooms, cleanliness, quality of the meals, medical care, ­in-­house activities, excursions, price, and so on. In a period of two weeks, about 30 questionnaires were collected from each respon-­ dent group. Some findings stood out:

Tools for Collecting Data about the Performance Shortcoming  117

• Meals were considered nutritional but rather boring, with little vari-­ ety from week to week, and this was one reason why quite a few residents ate less than they should. • The residents in poorest health found few recreational activities that suited their abilities, and those in good health said the same. It seemed the activities had been designed for an average that didn’t exist. • Relatives found the visiting hours to be too limited, often making it difficult to find time to come see their relatives. • Invoicing three months’ rent each quarter meant each bill was rather large and caused problems with payment. For many of these issues, changes could be made that would cost very little money. A more extensive meal rotation cycle was developed with the help of a local restaurant, the activities program was changed to accommodate the fitness levels of all residents, visiting hours were extended, and bills were issued monthly. A few months later, there were no vacant rooms, and both the residents and the relatives seemed much more content.

8.3  Check Sheet A check sheet is a table or form used for registering data as they are collected. One of the main applications is registering how often different problems or incidents occur. This provides important information about problem areas or probable causes of errors, and thus provides a good foundation for deciding where to concentrate during improvement. The approach is normally as follows:

1. Agree on what events are to be recorded. These must be clearly defined to avoid doubt whether an event truly occurred. It is usually also smart to include a category of “other” to capture incidents that are not easy to oth‑ erwise categorize.



2. Define the period for data recording and a suitable division into intervals.



3. Design the check sheet to be used during recording. An example is shown in Figure 8.1, where space is allocated for recording each event and summarizing both within the intervals and for the entire recording period.

118  Chapter Eight

Problem

Week 1

Week 2

Week 3

Total number of occurrences per problem

A

27

B C

4 23 53

D E F G

9 2 2 14

H Total number of problems per week

36

55

43

134

Figure 8.1  Example of a check sheet.



4. Perform data collection during the agreed period. In advance, it is necessary to make sure that everyone taking part in the data collection has a common understanding of the task, so as to achieve consistency in the data material.



5. When data collection is completed, analyze the material to identify events displaying a high number of occurrences. These will contribute to the prioritization of what specific areas within the chosen business process should be emphasized in the ensuing improvement work. A suitable visual aid for this analysis is a Pareto chart (see Figure 9.2).

The danger of overlooking one or more categories of problems should be men‑ tioned. If we are exclusively looking for events of the type defined in the check sheet, other problems might occur but fail to be recorded because the attention is directed only toward the expected problems. This is partly countered by the inclu‑ sion of an “other” category, but this issue should still be kept in mind.

Example A ­ medium-­sized company in the electrical installation industry annually submits a large number of bids for jobs for both private and industrial customers. The company was not satisfied with the portion of bids that resulted in jobs, and therefore wanted to follow up and improve the pro-­ cess of generating bids. To obtain an overview of the reasons why the company was not assigned more jobs, the check sheet shown in Figure 8.2 was designed. Each time a bid was rejected by a potential customer or the bid of a competitor was

Tools for Collecting Data about the Performance Shortcoming  119

preferred, the customer was asked to explain the reasons. The responses were entered into the check sheet during a period of three months. As the sheet indicates, price was not the dominating reason for rejec-­ tion, contrary to what the company had believed. Rather, the problem was very often a lack of flexibility with regard to when the job could be carried out. The result was that a new system for monitoring the availability of the service workers was designed. In addition, it was made clear in every sin-­ gle bid that the work could be started on very short notice and could very well be carried out in several separate periods, all at once, during evenings, or whenever. In the long run, this resulted in a radical increase in the portion of bids won, and at substantially better prices.

Cause of lost bid

January

February

March

Total number of occurrences per cause

Too high price

7

Poor quality Low flexibility

4 40

Poor impression during inspection

8

Low technical expertise

5

Total number of causes per month

23

20

21

64

Figure 8.2  Check sheet for recording the causes of lost bids.

8.4  Problem Concentration Diagram The check sheet presented in the previous section is a practical tool to collect data, especially about occurrences. When periods of time, areas of occurrence, or some other suitable headings have been defined for the columns in the check sheet, it can easily be used to record the desired data. However, in some cases, it is not that easy to label the columns properly. Say, for example, a rental car company is interested in determining where scratches and dents on its cars are most often located. To collect such data, a check sheet could perhaps be used, listing categories as hood, roof, trunk, different doors, and so on, as possible locations. However, the data could be too rough. Whenever a door has been indicated, there is no telling where on the door the damage was done. In such a case, a problem concentration diagram can offer more insight, and I am sure you have seen an example of such a diagram if you have ever rented a car.

120  Chapter Eight

When you pick up the car, the keys are typically in a small paper bag that contains fields for recording the time of return and the mileage (granted, there is a transition away from these ­paper-­based forms to electronic scanning upon rental car return, but the example is still valid), as well as a diagram of the car from different angles. If the car already has some damage, this is marked in the diagrams, and any new damages are duly recorded upon return. These diagrams can then be analyzed to see where scratches and dents show up most often, possibly as a basis for installing extra protection of these areas—for example, door protectors. It is possible to combine a check sheet with a problem concentration diagram. If the check sheet is suitable for data collection, the data recorded can be trans‑ ferred to a problem concentration diagram to be analyzed further. In this case, the diagram is perhaps more of an analysis tool that can be used when investigating the performance shortcoming in more detail. The steps in using problem concentration diagrams are the following:

1. Clearly define what events are to be recorded, and assign a symbol to each type of problem or event



2. Define the period for data recording and suitable intervals



3. Design the diagram by drawing a map of the system, object, or area



4. Populate the diagram with problem registration data through recording problem occurrences directly into the diagram

Example A ­medium-­sized city had been constantly criticized over the last few years for inadequate street lighting. People felt insecure in areas where the light-­ ing was poor, and the media claimed that both crimes and traffic accidents could be linked to dark or poorly lit areas of the city. The city’s technical operations department was in charge of the street-­ lights. It was equally frustrated with the problem, both because of the extra work from the constant repairs and because of all the negative attention to the situation. The paradox was that it had never changed as many light-­ bulbs or fixtures per month as it did in the last two years, effectively having increased productivity significantly by appointing a dedicated streetlight team with specialized vehicles at its disposal. There were extensive discussions about what caused so many bulbs or lamps to fail. Vandalism had been ruled out early on; only a few broken lamps were due to rock throwing or gunshots. In many cases, the bulbs simply went out much sooner than their estimated life expectancy; in other cases, the fixtures showed signs of burning. The bulbs were of the same brand, purchased from the same manufacturer as always, and seemed of consistent quality.

Tools for Collecting Data about the Performance Shortcoming  121

Truly puzzled by this enigma, the streetlight team decided to record every defective lamp that was fixed. This was done using a problem con-­ centration diagram in the shape of a large city map with all streetlights marked. All the repair teams had such a map in their cars, and whenever a lamp was fixed, it was marked on the map. The teams recorded some 400 defective lamps, and all lamps were displayed in the same map, with columns showing the number of lamps with defects in each area of the city. A limited and simplified cutout of this map is shown in Figure 8.3. This small part of the map may not show this very well, but the full version displayed extreme differences in concentration of defective lamps. Many areas had virtually no broken bulbs or fixtures, but others seemed to suffer from downright epidemics. At first, this gave the team no further clues as to what caused these anomalies, but almost by accident, an engi-­ neer from the local power company came upon the map. He immediately saw that the map quite accurately outlined the areas where new generators had been installed over the last two to three years. From tests the power company had undertaken, he knew that the voltage levels in these areas

Town hall Gas station

Stadium

Train station

= Defective lamps

Figure 8.3  Cutout of city map with streetlight problem concentration.

122  Chapter Eight

often ran higher than with the old generators. Further inquiries revealed that the fixtures used in the streetlights were vulnerable to elevated voltage levels, causing them to overload and break or the bulb to burn out. There was no ­short-­term solution to this problem, because there was no easy way of reducing the voltage delivered to the lamps. However, when fixtures were replaced, a new type was installed that was less sensitive. In the long run, this should reduce the problem.

The data available about the performance shortcoming, whether they are col‑ lected through the use of one of these tools or already existing, are now ready for analysis.

9 Tools for Analyzing the Performance Shortcoming

T

here are many possible tools that could be included in this section of the book. Therefore, I have selected a range from widely used, rather simple tools to less ­well-­known techniques:

• Critical incident • Pareto chart • Cause-and-effect chart • Five whys analysis • Scatter chart • Histogram • Relations diagram • Matrix diagram • Is–is not analysis • Fault tree analysis

9.1  Critical Incident Critical incident is a technique that can be used for identifying a process, sub‑ process, or problem that should be improved (Lawlor, 1985). After you’ve com‑ pleted this task and are in progress with an improvement project, it is also useful for figuring out what is causing the performance shortcoming.

123

124  Chapter Nine

It is a quite open and frank way of seeking information about organizational problems. The technique relies on honest responses from those involved, and a prerequisite is that the participants are completely free to express their views. In this respect, it is important that management display the right attitude to avoid censorship or withholding of information in fear of consequences from being too honest. The technique works as follows:

1. First, the participants in the analysis are selected, and this should include people actively involved in the business process to be improved.



2. Next, the group of participants is asked to answer questions such as the following: • Which incident last week was the most difficult to handle? • Which episode created the biggest problems in terms of maintaining customer satisfaction? • Which incident cost the most in terms of extra resources or direct expenditures?

The purpose is to focus on the critical incidents that, in one way or another, created problems for the employees, the organization, or other stakehold‑ ers. The period covered by the questions can range from a few days to several months. It is, however, not favorable if the period is too long, as it might be difficult to determine the most critical incident simply because many incidents can qualify as candidates given a sufficient length of time.

3. The collected answers are sorted and analyzed according to the number of times the different incidents have been mentioned. A graphical repre‑ sentation format might very well be used for this purpose. The incidents occurring most often, such as the critical ones, are obvious candidates for prevention of recurrence.

Example A large corporation started a project to improve customer service in its staff of 15 switchboard operators. It was difficult to decide where in this area to start, so it was decided to try the technique of critical incident. All the operators were asked to describe the most difficult situations they had experienced the last month. The resulting events were sorted by frequency and presented in a chart, as shown in Figure 9.1. As the

Tools for Analyzing the Performance Shortcoming  125

chart shows, the most critical incidents were “the wanted person does not answer” and “the operator does not know who should handle the request.” This resulted in efforts focused around designing a system to keep track of the individual employee and drawing up clearer rules for who should handle which requests. Number of occurrences 18 15

14

10

6 5

3

Other employees interrupt by coming into the office

Calls that have been directed are not transferred to the right extension

The call disappears during transfer

The operator does not know who should handle the request

The wanted person does not answer, and no absent note has been entered

4

Incident

Figure 9.1  Critical incidents for the switchboard operators.

9.2  Pareto Chart The Pareto chart is based on the Pareto principle, formulated by the Italian mathematician Vilfredo Pareto during the 1800s. He was concerned with the distribution of riches in society and claimed that 20 percent of the population owned 80 percent of the wealth. Translated to modern quality terminology, the Pareto principle states that most of the effects, often around 80 percent, are from a small number of causes, often only 20 percent. For example, usu‑ ally 80 percent of the problems related to purchased material are caused by

126  Chapter Nine

20 percent of the suppliers. Even more importantly, 80 percent of all costs connected to poor quality or generally low performance is from 20 percent of all possible causes. A healthy approach is therefore to start the improvement work by attacking these 20 percent, often labeled “the vital few.” This does not imply that the remaining 80 percent should be ignored; “these important many” should in due time also be addressed. The Pareto principle only says something about the order in which the problems should be attacked. The Pareto chart is a tool used to graphically display this skewed distribu‑ tion, the 80–20 rule. The chart shows the causes of a problem sorted by the degree of seriousness and expressed as frequency of occurrence, costs, perfor‑ mance level, and so on. The causes are sorted by placing the most severe on the left side of the chart, rendering it quite easy to identify the vital few. To enable portraying additional information in the chart, it is common to include a curve showing cumulative importance. This is depicted in the general chart in Figure 9.2. The following are the most common steps in using a Pareto chart:

1. Define the performance shortcoming and the potential causes of it.



2. Decide which quantitative measure to use when comparing the possible causes. As has been mentioned, this measure might be how often the dif‑ ferent problems occur or consequences of them in terms of cost or other conditions.



3. Use existing data or collect the necessary data.



4. Place the causes from left to right along the horizontal axis in descending rela‑ tive importance. Draw rectangles of heights that represent this importance.

Problems

100%

50%

A

B

C

D

E

F

Figure 9.2  A general Pareto chart with a line for cumulative importance.

Tools for Analyzing the Performance Shortcoming  127



5. Mark the data value on the left vertical axis and the percentage value on the right, and draw the curve for cumulative importance along the top edge of the rectangles.

A quick inspection of the chart can now answer questions like, what are the two or three main causes of the low performance level of this process? or how large a portion of the costs can be attributed to the most vital causes? This information can be used for actively directing the improvement efforts toward areas that are likely to produce the best effects. Pareto charts can easily be constructed using standard spreadsheet software.

Example The installation company in the check sheets example also drew a Pareto chart displaying the collected data. Instead of how often different reasons were given for why the company’s bid did not win, the value of the lost job was assigned to the vertical axis. The resulting chart is shown in Figure 9.3. The chart confirmed that the low flexibility expressed to the customers regarding how quickly and at what time of day the work could be done not only was the most frequent reason for failing to win bids, but also caused the biggest losses in terms of ­lost-­job value. It was, in other words, right to aim for a new system for planning the work in the company. Lost-job value 100%

150,000

50%

Figure 9.3  Pareto chart for value of lost jobs.

Low technical expertise

Poor quality

Poor impression during inspection

High price

Low flexibility

75,000

128  Chapter Nine

9.3 ­ Cause-­and-Effect Chart The cause-and-effect chart is one of the classical and most widely used tools in quality management, and not without reason. Many view the tool as being some‑ what ­old-­fashioned and sturdy in a rather boring sense, but it is truly a useful tool with many powers. The main purpose is, as the name implies, to identify possible causes of an effect. The effect being analyzed can be an experienced problem or a future ­hoped-­for state where the causes no longer occur. There are at least two types of ­cause-­and-effect charts: • Fishbone chart • Process chart The fishbone chart is the traditional way of constructing such charts, where the main product is a chart whose shape resembles a fishbone. The principles for this chart are shown in Figure 9.4. There are two ways of creating the chart: • Dispersion analysis, where the effect being analyzed is drawn on the right side of the chart, at the end of a large arrow. Main groups of probable causes are drawn as branches to the arrow. For each branch, all possible causes are identified. • Cause enumeration, where all probable causes are simply brainstormed and listed in the order they were generated. When this has been done, the causes are grouped and drawn in the fishbone chart. The end product is the same regardless of the approach.

Main category

Main category

Cause

Cause Cause Cause

Cause Cause Cause

Cause Cause Main category

Cause Cause Cause

Main category

Figure 9.4  The structure for a fishbone chart.

Problem to be solved (effect)

Tools for Analyzing the Performance Shortcoming  129

The main steps for producing a fishbone chart using dispersion analysis, which is the more common approach, are the following:

1. Assemble a suitable group of individuals who possess the necessary knowledge about the area to be analyzed.



2. Clearly describe the effect for which causes are sought. This effect is often a low performance level for one of the business processes of the organization.



3. Using a white board or some other large medium, draw the effect at the end of a large arrow. The point is to set aside enough space for the gener‑ ated causes, not symmetry or nice drawing effects.



4. Identify the main categories of possible causes of the effect and place these at branches emanating from the large arrow. For physical processes, some common main categories are: • People • Machines and equipment • Materials • Methods • Measures • Environment—culture, organizational structure, physical environment, and so on Similarly, for service processes, the traditional categories are: • People • Processes • Framework conditions • Work environment



5. Brainstorm all possible causes and place these in the suitable area of the chart. Emphasize brief and succinct descriptions. Proceed through the chart one main category at a time, but also include suggestions that belong to categories other than the one currently being treated. Causes that belong to more than one category are placed in all relevant positions. It is often required to redraw the chart after the first version has been completed.



6. Analyze the identified causes to determine those that should be addressed fur‑ ther. Remember that the purpose is to cure the problem, not the symptoms.

Step 1

Cause

Cause

Cause

Cause

Main category

Cause

Cause

Cause

Cause

Main category

Cause

Cause

Cause

Cause

Cause

Step 4

Problem to be solved (effect)

Cause

Cause Main category

Cause Main category

Cause Cause

Cause Cause

Main category

Cause

Cause

Cause

Cause

Main category

Main category

Problem to be solved (effect)

Step 3

Cause Main category

Cause Cause

Cause Cause

Main category

Cause

Cause Cause

Main category

Cause Cause

Cause

Cause

Step 2 Main category Cause

Cause

Problem to be solved (effect)

Main category

Main category

Cause

Cause

Cause

Figure 9.5  Process chart with fishbone charts for each process step.

Start

Cause

Cause

Main category

Cause

Cause

Cause

Cause

Main category

Main category

Cause

Problem to be solved (effect)

Finish

130  Chapter Nine

Tools for Analyzing the Performance Shortcoming  131

The process chart variant of the fishbone chart is more directly aimed at improve‑ ment of business processes. The main steps in the process to be improved are drawn. For each step of the process believed to create problems or contribute to a low overall performance level, a fishbone chart is constructed that deals with all potential causes of why this step performs less than expected. For each of these individual fishbone charts, the process is the same as outlined earlier for construct‑ ing such a chart. After individual charts for each problem step of the process are designed, a collective analysis is undertaken to identify the causes that seem to be of the high‑ est importance. For these, solutions are sought that can reduce the negative effects on the overall performance level of the process. The resulting chart might look like the one shown in Figure 9.5.

Example A pump manufacturer experienced frequent defects in a series of pumps delivered to a customer. Closer inspection revealed that most of the defects stemmed from inaccurate dimensions of two shafts used in the pumps. A group consisting of designers, the manufacturing manager, the shaft department manager, and several operators was established. The objec-­ tive was to find and eliminate the causes of the problems. The design of the fishbone chart in Figure 9.6 gave some clues as to where the problems might be found. From the fishbone chart, it was evident that the physical environment where the shafts were manufactured was rather unsuitable for the process, as a low temperature in the shop along with the wrong coolant resulted in most of the dimension errors.

Employees Low motivation

Methods Unsuitable process route

Poor training

Unsuitable hardening Wrong coolant

Poor maintenance

Often too low temperature Environment

Equipment

Figure 9.6  Fishbone chart for shaft manufacturing.

Dimension errors Too infrequent tool sharpening

132  Chapter Nine

9.4  Five Whys Analysis The five whys analysis is also known as the ­why-­why chart and root cause analy‑ sis. As these names imply, the purpose is to find the true root cause of a problem (Andersen and Pettersen, 1996). The technique can very well be used in connec‑ tion with a ­cause-­and-effect chart to analyze each identified cause to ensure that it really is the root cause of the problem and not only a symptom of another and more deeply rooted cause. This can in fact be compared to peeling an onion, where each layer is removed to reveal another layer, until the center of the onion is reached. The procedure for conducting a five whys analysis is as follows:

1. Determine the starting point, either a problem or a ­high-­level cause that should be further analyzed.



2. Use brainstorming to find causes at the level below the starting point.



3. For each identified cause, pose the question, why is this a cause for the original problem?



4. For each new answer to the question, ask the question again and again until no new answers result. This will probably be one of the root causes of the problem. As a rule of thumb, this often requires five rounds of the question why.

If the question is slightly altered to ask “how” instead of “why,” the technique can be used to find root means for reaching a desired state or effect. The analysis can be conducted in different ways. A graphically visible way of keeping track of the different levels of causes is listing them below one another, as shown in Figure 9.7. In this example, a manufacturing company is attempting to reduce the amount of work in progress. The result of the analysis is that the key to reducing the work in progress is to develop good relationships with the suppliers. If this analysis had Low level of work in progress Why?

Maintain no stock of finished goods

Why?

Short manufacturing time

Why?

Run small batch sizes

Why?

Frequent and swift deliveries from the suppliers

Why?

Extremely good relationships with the suppliers

Figure 9.7  List representation of five whys.

Tools for Analyzing the Performance Shortcoming  133

not been undertaken, the company might be led to believe that the answer was simply to remove the finished goods inventory, which could have had serious con‑ sequences. Alternatively, a diagram might be used to portray an entire network of causes at different levels, as shown in the following example.

Example A video rental chain had experienced an ­ever-­increasing degree of customer dissatisfaction when measuring customer satisfaction data. A simple questionnaire revealed the four most important causes: • Long checkout times • Poor selection of titles • Impolite and unfriendly personnel • Unfavorable store location and layout Through a five whys analysis, the chart in Figure 9.8 was constructed. Most of the problems could thus be attributed to low wages and low management expertise.

Manual files Long checkout times

Dissatisfied customers

Poor title selection

Impolite personnel

Poor store location and layout

Figure 9.8  Five whys analysis chart.

No computer system

Low management expertise

Poor employee training

High employee turnover

Low wages

Poor customer analysis

Poor question and selection plan

Low management expertise

Poor employee training

High employee turnover

Low wages

Poor employee training

High employee turnover

Low wages

Manual invoicing

No market analysis performed No competitor analysis performed

Low management expertise

134  Chapter Nine

9.5  Scatter Chart A scatter chart can be used to show the relationship between two variables. The variables can be process characteristics, performance measures, or other condi‑ tions and are usually measured at specified time intervals. When one of the factors increases, the other can also increase, decrease, or display only random variation. If the two variables seem to change in synchronization, it might mean that they are related and impact each other. For example, we might find that the number of defects increases in proportion with the amount of overtime. However, keep in mind that even if there is some degree of synchronized variation between variables, it does not mean for certain that there is a ­cause-­and-effect relationship between them; it may very well be a third variable that is causing the effects. This can be illustrated by the fact that a scatter chart was once constructed that showed an obvi‑ ous correlation between the Dow Jones index and the water level of Lake Superior in the years between 1925 and 1965. Y

Y

Y

X

X

Strong positive correlation

Possible positive correlation Y

Y

X No correlation

Strong negative correlation

X

Figure 9.9  Scatter charts showing different degrees of correlation.

Possible negative correlation

X

Tools for Analyzing the Performance Shortcoming  135

The degree of relationship between the variables being examined can range from highly positive to highly negative correlation. Between these two extremes are weaker degrees of both positive and negative correlation, as well as no correlation. Figure 9.9 shows some examples of scatter charts for different degrees of correlation. When using a scatter chart, the following steps are usually conducted:

1. Select the two variables, one independent and one dependent, to be examined.



2. For each value of the independent variable, the corresponding value of the dependent variable is measured. These two values form a data pair to be plotted in the chart. Typically, there should be at least 30, but preferably more than 100, data pairs to produce a meaningful chart.



3. Draw the chart by placing the independent variable, that is, the expected cause variable, on the x axis, and the dependent, the expected effect vari‑ able, on the y axis.



4. Plot the collected data pairs on the chart and analyze them. If the chart shows no correlation, the data pairs can be drawn in a logarithmic chart. Such a chart can reveal connections that are not visible in a chart with ordinary axes.

Producing a scatter chart by hand, at least with some amount of data, can be tedious work. Spreadsheet software does the job much more easily. Remember, if the chart indicates a relationship, the variables should be further analyzed to confirm this. In the same way, a chart showing no relationship should not automatically lead to the conclusion that no relationship exists.

Example As part of efforts to create an improved state of readiness for performing rush jobs, the electric installation company we looked at earlier wanted to examine the relationship between the amount of jobs—and thus the need for ready servicemen—and the number of days with thunder and lightning. Through a quarter of a year, the company every week counted the num-­ ber of jobs and the number of days with thunder and lightning. The scat-­ ter chart that was designed (see Figure 9.10) displays a clear correlation between bad weather and the amount of jobs. In response, the company started considering the weather forecasts when planning capacity and the possibility for reserve capacity.

136  Chapter Nine

Amount of jobs 25

20

15

10

5

0

1

2

3

4

5

6

Number of days with thunder and lightning

Figure 9.10  Scatter chart to analyze weather and jobs.

9.6  Histogram A histogram, or a bar chart as it is also called, is used to display the distribution and variation of a measure of some sort. The measure can be an infinite number of differ‑ ent aspects—for example, length, diameter, duration, and costs. The same informa‑ tion can also be presented in a table, but it can be difficult to spot important patterns in the data material. The graphic presentation format makes it easier to see relationships. Unlike the Pareto chart, which is really a variant of a histogram that displays the fre‑ quency of occurrence for a phenomenon, a histogram portrays intervals of data. The following list outlines the construction of a histogram. The procedure is illustrated using an example where the diameter of a hole has been measured for 125 work pieces. The data material is shown in Table 9.1.

1. Count the number of data points in the data material and label this number N. To produce a valid histogram, there should be at least 50 data points. In the example, N = 125.



2. Determine the numerical distance R between the largest and smallest value in the data material. In the example, R = 10.7 – 9.0 = 1.7.



3. Depending on the number of data points, R is divided into a number of equally large classes, C. The number of classes can be found in Table 9.2. For N = 125 in the example, the number of classes should be between 7 and 12. We have decided to use 10 classes, that is, C = 10.

Tools for Analyzing the Performance Shortcoming  137

Table 9.1  Measures of the diameter of a hole for 125 work pieces. 9.9

9.9

10.4

9.8

10.1

10.2

9.8

10.3

9.7

9.7

9.8

9.3

10.2

9.3

9.2

9.8

9.8

10.1

9.8

9.8

9.7

9.8

10.2

9.8

10.2

10.0

9.7

9.5

9.6

9.5

10.2

9.4

10.1

10.1

9.6

9.7

10.0

10.0

9.3

9.5

9.9

10.1

9.6

9.7

9.6

9.5

9.7

9.7

10.0

9.7

9.3

10.7

9.8

9.8

9.8

9.9

9.6

9.7

9.7

9.9

9.0

10.2

9.3

10.3

9.9

9.9

10.1

10.7

10.7

9.6

10.0

9.5

9.2

9.9

10.0

10.1

10.0

9.8

9.4

9.3

9.5

9.7

9.7

9.7

9.8

10.2

10.4

9.6

9.9

9.6

9.7

9.4

9.8

9.9

10.3

9.8

10.0

10.0

10.3

9.4

10.6

9.4

9.8

9.8

9.5

10.7

10.1

9.5

9.6

10.1

10.1

9.6

9.3

9.5

9.9

10.3

9.9

9.5

9.7

10.1

10.0

10.0

9.6

9.4

9.9

Table 9.2  Determining the number of classes, C. Number of data points, N





Number of classes, C



Fewer than 50



5–7



50–100



6–10



100–250



7–12



More than 250

10–20

4. Determine the width of each class, labeled H. This is calculated using the following formula: H=

R 1.7 = = 0.17 ≈ 0.2. C 10

As can be seen, the class width in the example is rounded off to 0.2, as the width should always have the same number of decimal places as the data points—that is, one decimal place in this case.

5. Determine the lower and upper values for the individual classes. This is done by letting the smallest value in the data material become the lower value for the first class. The upper value for this class is then found by adding the class width to the lower value. In the example, the first class will span the range from 9.0 to 9.2. The next class starts at 9.2 and covers the values up to 9.4, and so on. Remember that the lower value always

138  Chapter Nine

includes this value, while the upper value only goes up to it. The value 9.2, in other words, belongs to the second class, not the first one.

6. To simplify the construction of the histogram, the data material is inserted into a check sheet. For the example, the check sheet is shown in Figure 9.11.



7. Finally, the histogram is constructed according to the check sheet. The classes are marked along the horizontal axis and the frequency along the vertical. The distribution among the classes is indicated by bars. The resulting histogram for the example is shown in Figure 9.12.

Frequency

Total

Class

Lower value

Upper value

1

9.0

9.2

1

2 3

9.2 9.4

9.4 9.6

4 5 6 7

9.6 9.8 10.0 10.2

9.8 10.0 10.2 10.4

9 16 27

8

10.4

10.6

12 2

9

10.6

10.8

4

10

10.8

11.0

0

31 23

Figure 9.11  Check sheet for the example. Frequency 35 30 25 20 15 10 5 9.0 9.2 9.4 9.6 9.8 10.0 10.2 10.4 10.6

Figure 9.12  Histogram for the example.

Class

Tools for Analyzing the Performance Shortcoming  139

Example Imagine that this histogram was designed by a company manufacturing steel components that should have a hole with a diameter between 7.5 and 10.5 millimeters. With the help of the diagram, the company can see that the process falls beyond the upper specification limit and will actually manufacture approximately 3 percent defects. This information can thus be used to adjust the process to avoid defects.

Using histograms this way is actually highly connected to statistical process con‑ trol, which is treated in section 11.4. Whereas a control chart in statistical process control contributes to a continuous monitoring of whether a process is under con‑ trol, a histogram can reveal permanent deviations that do not appear in a control chart. The shape of the histogram should be closely inspected and interpreted to reveal problems in the process. Ideally, a histogram will produce an image of the variation in the data mate‑ rial while at the same time providing a sensible level of detail. If, however, too few classes have been defined, only a few bars will be the result, and these are not suited for revealing any patterns. Correspondingly, too many classes will hide the pattern because some will be empty, which results in an appearance resembling a comb. Particular patterns in the histogram indicate some typical problems in the process: • A pattern with one obvious peak shows the mean value for the process. Depending on how much the process varies around this peak, the process is defined to be good or bad. A diagram like the one shown in Figure 9.13 is a process displaying a small variation width that lies centered within the spec‑ ification limits. This is a good process. The diagram in Figure 9.14 shows a poorer process with a large variation width. If this cannot be reduced, the process will produce results outside the limits, which can be countered by conducting a 100 percent control. • Furthermore, the location of the peak in diagrams with only one peak deter‑ mines whether the process is under control. As was mentioned, the pro‑ cess in Figure 9.13 is a good process displaying little variation and lying well centered within the limits. In Figure 9.15, another process with little variation is shown, but this one is not centered. Adjusting this process could result in a good process.

140  Chapter Nine

Frequency Lower specification limit

Upper specification limit

Class

Figure 9.13  Centered process with a small variation width. Frequency Lower specification limit

Upper specification limit

Class

Figure 9.14  Centered process with a large variation width.

• When the histogram displays two clear peaks, as shown in Figure 9.16, this can be the result of different causes. Either the data points can stem from two different sources, which should be checked, or the process’s mean value has changed during the data collection interval. • A ­cut-­off pattern—that is, a histogram ending abruptly and showing no signs of tapering off—is a sign of control or selection of the results. This is shown in Figure 9.17 and is often visible close to the tolerance lim‑ its. In this case, a control has probably been conducted, and results fall‑

Tools for Analyzing the Performance Shortcoming  141

Frequency Lower specification limit

Upper specification limit

Class

Figure 9.15  Process with a small variation width, but not centered. Frequency

Class

Figure 9.16  Histogram with two peaks.

ing outside the limits have been removed from the data material. In these cases, we must know how much the defects cost. The process should be improved to avoid the defects. • The previously mentioned comblike pattern is an indication that too many classes have been defined. However, such a pattern can also be a result of problems with the measurement equipment. If the measurement equipment is unable to register points in the individual classes or with the accuracy required to use so many classes, such an appearance will be the result. In such a case, the measurement approach should be reevaluated.

142  Chapter Nine

Frequency Lower specification limit

Upper specification limit

Class

Figure 9.17  Histogram that is cut off.

9.7  Relations Diagram A relations diagram is a tool to identify logical ­cause-­and-effect relationships in a complex and confusing problem or situation. For problems or situations where there is a network of such relationships, a relations diagram is particularly useful, as it has the ability to visualize them. There are two types of relations diagrams: • Qualitative relations diagram • Quantitative relations diagram In the qualitative diagram, both the problem and the causes at several levels can be included, as shown in Figure 9.18. The diagram is actually quite simi‑

Cause

Main cause

Cause

Cause

Cause

Problem

Cause

Cause

Main cause

Cause-and-effect relationship

Figure 9.18  The principles of a qualitative relations diagram.

Tools for Analyzing the Performance Shortcoming  143

lar to the traditional ­cause-­and-effect chart, but it is more suited for complex problems. To generate a qualitative relations diagram, follow these steps:

1. Isolate all factors believed to be related to the problem. Without forming an opinion about the relationships between the factors, each of these is freely expressed on an individual basis. Boxes can very well be drawn that contain the factors.



2. Identify the causal relationships between the factors and illustrate these with the help of arrows in the diagram.



3. Classify the factors depending on which role they play in the ­cause-­andeffect situation.



4. Concentrate the improvement effort around the main causes of the problem.

Example After having spent much time and money introducing a system for perfor-­ mance measurement, a company learned that the system was not widely used and was viewed with little respect among the employees, at times even directly sabotaged. The company spent quite some time construct-­ ing the relations diagram in Figure 9.19, and it found two causes of the problem.

Inaccurate measurement

Lacking measurement training

Poor knowledge about data collection

Nonmotivating measures

The measurement system does not work

The measures are not taken seriously

Poorly defined measures

Some measures conflict

Figure 9.19  Qualitative relations diagram for a poorly functioning measurement system.

A quantitative relations diagram can sometimes be easier for determining the role of different factors. In Figure 9.20, a general diagram is shown, constructed according to the following approach:

144  Chapter Nine



1. Place the factors to be included in the analysis throughout the diagram, preferably in a coarse circular shape.



2. For each factor, assess which other factors this impacts or is impacted by, and indicate these impacts with arrows. The direction of the arrow indi‑ cates the direction of the impact—that is, an arrow pointing to factor B from factor A means that factor A impacts factor B.



3. After all relationships have been assessed, the number of arrows is counted and denoted in the diagram.

Depending on the number of arrows pointing in each direction for a factor, the fac‑ tor can be defined to play one of two roles: • Performance driver or cause—that is, a factor that impacts or enables the performance level of another factor. Another term is throttle vari‑ able. A performance indicator has more arrows pointing away from it than to it. • Result indicator or effect—that is, a factor that indicates a consequence as a result of a performance driver. A result indicator has more arrows pointing to it than away from it. When trying to find the main cause of a problem or an effect, ultimately wanting to identify the factor that will increase the performance if it is improved, the perfor‑ mance drivers must form the starting point. They drive the process and thus create its performance level.

0 in 4 out 2 in 0 out

Factor A

Factor E Factor B 1 in 3 out

Factor D 2 in 0 out

Factor C 2 in 0 out

Figure 9.20  A general quantitative relations diagram.

Tools for Analyzing the Performance Shortcoming  145

Example If a quantitative relations diagram were used in the preceding example instead of a qualitative one, the result would be as depicted in Figure 9.21. This diagram gives the same information: the measurement system does not work (7 in, 0 out), while the main performance drivers are poorly defined measures (0 in, 5 out) and lacking measurement training (0 in, 3 out). 0 in 3 out

3 in 2 out

Lacking measurement training 1 in 2 out

Inaccurate measurement

Poor knowledge about data collection

1 in 1 out

3 in 1 out

Nonmotivating measures

The measures are not taken seriously

Poorly defined measures

The measurement system does not work

0 in 5 out

7 in 0 out Some measures conflict 1 in 2 out

Figure 9.21  Quantitative relations diagram for the example.

9.8  Matrix Diagram So far, we have looked at tools that, in different ways, contribute to identifying relationships between different factors, often in the shape of ­cause-­and-effect rela‑ tionships. The matrix diagram has about the same objective, but the force of this tool compared with many of the others lies in its ability to graphically portray the strength of such relationships. Like many of the other tools, this one can also be used in several stages of improvement work, such as prioritizing improvement areas, identifying problems and causes, and planning.

146  Chapter Nine

Depending on the shape of the matrix, and thus the number of variables ana‑ lyzed, there are several types of matrix diagrams, as shown in Figure 9.22 (Swan‑ son, 1995): • Roof-shaped, which was briefly mentioned in section 5.6. Unlike the other types of matrix diagrams, where the strength of the relationships is given in only one direction, the roof matrix classifies relationships as neutral, posi‑ tive, or negative. • L-shaped • T-shaped • Y-shaped • X-shaped • In addition, there is a C-shaped matrix that treats ­three-­dimensional rela‑ tionships, but is rarely used due to its complexity. The number of variables analyzed in the different formats, as well as the number of direct and indirect (that is, through a third variable) relations, is summarized in Table 9.3.

L-shaped

Variable 2

Variable 1

ia bl e

Va r

Va r

Variable 2

Variable 1

2

Roof-shaped

Variable 1

Variable 1

Variable 2

Variable 3 Variable 4

Variable 3

Y-shaped

T-shaped

Figure 9.22  Types of matrix diagrams.

X-shaped

ia

bl e

3

Tools for Analyzing the Performance Shortcoming  147

Table 9.3  Formats with the number of variables and relations for matrix diagrams. Matrix shape

Number of variables

Direct relations

Indirect relations

L

2

1

0

T

3

2

1

Y

3

3

0

X

4

4

2

C

3

3 simultaneously

0

Roof

1

—

—

As explained in section 5.6, a standard set of symbols is used for indicating the strength of relations between variables. Figure 9.23 shows these symbols and the corresponding weight factors. The steps when using a matrix diagram are the following:

1. Select the variables to be analyzed for potential relationships.



2. Select the matrix format according to the number of variables and the number of expected relations. Table 9.3 can aid in this task.



3. Insert the variables into the diagram.



4. Indicate relations by using the symbols in Figure 9.23. Do not be tempted to use corresponding weight factors inside the actual matrix, as this reduces the readability of the diagram.



5. For each row and column in the diagram, the corresponding weight factors for each relations symbol are summarized and the sums are presented in the diagram.



6. Variables with a high sum play an important role in the situation and should be analyzed further.

Relation

Symbol

Weight

Weak

1

Medium

3

Strong

9

Figure 9.23  Relations symbols.

148  Chapter Nine

Example The video rental chain described in a previous example wanted to assess the impact of the different business processes with regard to customer needs and expectations. The analysis was performed in an L-shaped matrix diagram, as shown in Figure 9.24. With the diagram, the company could answer a number of questions. If, for example, it wanted to improve the checkout time, the relevant pro-­ cesses would be the checkout process itself as well as the information system used. Low prices, on the other hand, could be achieved through improvements in almost all processes.

Processes Customer expectations

Market Title Competitor Information analysis Checkout selection Advertising analysis system Training Total

Low prices

22

Good selection

34

Several copies

21

Store layout

12

Store location

3

Speedy checkout

21

Friendly staff

12

Total

18

22

21

6

16

18

24

125

Figure 9.24  Matrix diagram for the video rental chain.

9.9  Is–Is Not Analysis One phenomenon I have observed quite a few times in improvement projects is an idea overload. Many who regularly work with improvements become quite adept at taking brainstorming sessions to the fullest, and, as a result, many teams find themselves inundated with ideas about causes of low performance or possible improvements that can be made. In such cases, the danger is that ideas are so plen‑ tiful it becomes difficult to separate the essential from the trivial. The is–is not analysis is a tool that helps you see distinctions and clarify what the problem is not about. Using this tool can help you understand what is most likely to cause the performance shortcoming as well as identify which issues are definitely not related to it. Even at a later stage in the improvement cycle, when many ideas for improvement actions are on the table, the analy‑

Tools for Analyzing the Performance Shortcoming  149

sis can be used to distinguish between what is and what is not related to the shortcoming. The steps to undertake an is–is not analysis are:

1. Produce an empty matrix of six rows by four columns and fill the header column with the standard headings shown in Figure 9.25.



2. In the ­upper-­left corner of the matrix, state the performance shortcoming being analyzed.



3. Fill in the second column with “is” information: what, where, when, who, and so on.



4. In the same manner, fill in the third column with “is not” information.



5. Compare the two columns for anything odd or that stands out and place these in the fourth column.



6. For each element in the fourth column, analyze how it could be a cause of the performance shortcoming.



7. For the possible causes identified this way, test them by checking if they explain all items in the “is” and “is not” columns. Those that do are likely the real cause(s).

Performance shortcoming

Is

What occurs, what objects are affected? Where does the problem occur? When does the problem occur? Extent of shortcomings?

Who is involved?

Figure 9.25  Generic table for is–is not analysis.

Is not

Distinction

150  Chapter Nine

Example A large ski resort operates a number of different ­sub-­businesses: ski lifts, ski equipment rental, a ski store, restaurants, and a hotel. Lately, several accidents have been reported in connection with the ski lifts. The regulat-­ ing body granting licenses to operate the lifts was notified, and following an inspection, the resort was given a month to investigate the incidents and come up with an action plan to prevent further accidents. The matrix in Figure 9.26 emerged from the analysis of the reported incidents that occurred during the previous year. The analysis gave clear indications that the injuries occurred only in special areas of the resort and were caused by a certain group of guests. After further investigation, it was discovered that the tree barrier that was supposed to keep skiers away from the lift area had been compromised at both lifts 2 and 5. A couple of trees had fallen, and branches appeared to have been cut by a knife. This was enough to create an opening that allowed a shortcut between different slopes. Normally one would have to ski to the bottom, traverse a flat area to the next lift, and take this to the

Performance shortcoming

Is

Is not

Distinction

What occurs, what objects are affected?

People on the ground hurt by objects dropped from the lifts

No one has been hurt by falling out of the lifts

Injuries occur as a result of objects falling from the lifts

Where does the problem occur?

During the top 50 yards of lifts 2 and 5

Lifts 1, 3, and 4 or lower parts of lifts 2 and 5

Only a small part of the lift area seems to be affected

When does the problem occur?

95% of the incidents occurred during weekends

Very few injuries during weekdays

Weekends seem to be the problematic time period

Extent of shortcomings?

Injured guests between 7 and 22 years old

No injuries among adults older than 22

Injuries seem to afflict younger guests

Who is involved?

Objects (skis, helmets, poles) dropped by inexperienced skiers

Only one case of an experienced skier dropping a ski

Inexperienced skiers with rented equipment cause most incidents

Figure 9.26  Ski resort is–is not analysis.

Tools for Analyzing the Performance Shortcoming  151

top. Putting up a railing or some other physical barrier would block this shortcut and eliminate most of the problem. The other observation was that only inexperienced skiers—more spe-­ cifically, people who rented equipment from the resort—dropped skis and other objects from the lifts. Such objects could not hurt other guests after the shortcuts had been closed, but this was still annoying, both to the guests and to the staff who had to retrieve the objects. Since service per-­ sonnel on the slope had written down the names of some of the people involved in the accidents, three of the people who had dropped skis could be interviewed. They all complained that they received very little training in how to put on the boots, skis, and so on, properly when renting the gear. As a result, the boots were loose on their feet and were not properly attached to the bindings. The process for equipment rental was changed: one attendant picked out the correct equipment, and another attendant explained to the guests how to use the gear properly.

9.10  Bottleneck Analysis Bottleneck analysis derives from a large body of research contained within the optimized production technology (OPT) theory and the theory of constraints, developed by Goldratt (see both Goldratt and Cox, 1992, and Goldratt, 1990). One of the key messages of these theories revolves around bottlenecks in production systems (where a production system can deliver products or services). A produc‑ tion system consists of a number of resources that are used to transform inputs to outputs. Resources can be people, computers, machines, cars, and so on, and they can all be classified as one of two types: • Bottleneck, meaning a resource with a capacity that is less than or equal to the demand for the resource from the market • Non-bottleneck, which has a capacity greater than the demand Metaphorically speaking, a bottleneck can be seen as a constriction or narrowing of the flow passage in a production system or business process, that is, a point in the process that limits the flow and reduces it to a narrow stream:

Bottleneck

152  Chapter Nine

A simple example illustrates this, where the process to manufacture a prod‑ uct consists of five steps, as shown in Figure 9.27. At a demand of 2000 units per week, the process can deliver the products with its current capacity. If demand is increased to 2200 units in a week, step three in the process has too low of a capacity to keep up with demand and becomes the bottleneck. Although the other four machines can take the increased demand, the output from the entire process will be limited by this one bottleneck resource. The bottleneck is the weakest link in the chain and controls how much the whole process is able to produce. The theory of constraints contains much discussion about the properties of bottlenecks and how bottleneck and ­non-­bottleneck resources are linked. I will not go into detail here, but I hope you understand the significance of a bottleneck as a limiting factor in the performance of a process. From this it follows that knowing which resources constitute bottlenecks is useful. First, you can eliminate a bottleneck by increasing its capacity or redesign‑ ing the process to avoid using the bottleneck. This will increase the capacity of the process but only up to the point where another resource takes over as the bottleneck; bottleneck elimination is a continual battle. Second, identifying the bottleneck allows you to optimize your processes to maximize the conditions for this resource. You can divert production from the bottleneck to some other resource to free up time. If the bottleneck is a machine or some other type of physical equipment, you should make sure maintenance is carried out according to plan. If a human is the bottleneck, take good care of this employee to ensure that she or he stays healthy, enjoys the job, and so on. Further, always make sure the bottleneck resource is kept going, perhaps by keeping a buffer stock in front of it. And since an hour lost in a bottleneck is an hour lost for the entire process (according to OPT rules), make sure you don’t allow the bottleneck to produce defective parts or services. Capacity

Process flow

1

2

3

4

5

Figure 9.27  A bottleneck resource limits the flow from the entire process.

Tools for Analyzing the Performance Shortcoming  153

Nevertheless, how you deal with a bottleneck once it has been identified is not an integral part of the analysis. The purpose of the bottleneck analysis is to identify the bottleneck(s) of the process or production system. The following steps detail how to identify a bottleneck:

1. Model the business process or production system, using either a flowchart or a network diagram.



2. Identify the demand for products or services from the process or produc‑ tion system.



3. Start from the end of the process and determine the capacity need for each resource by multiplying the volume demanded by the cycle time in each resource.



4. Compare the capacity need with the available capacity, preferably by showing these as bar charts with two bars for each resource in the flow‑ chart or network diagram. Resources where the capacity need exceeds the available capacity are bottlenecks.

Notice from the definition of a bottleneck and how the analysis is carried out that a bottleneck is only momentary. If demand composition is changed, the bottleneck could move to another resource or the process could function without any bottle‑ necks. However, for stable processes or production systems with a similar demand composition over time, the bottleneck often stays the same.

Example A public university hospital serves as an emergency ward for its town, meaning the doctors will see anything from serious emergency victims to ­after-­hours patients with bruised knees or sprained fingers. During a normal week, especially between 7 and 11 p.m., a line of waiting patients would form, many of who were in pain from accidents. Some complained loudly, and others waited in silence. The line created frustration and stress for both patients and personnel. Some months before, the entire hospital had gone through a project of mapping its processes, including the emergency ward. The flowcharts showed a set of processes that sometimes ran through many units/types of personnel, with the typical patient following this route: • Immediately after entering the emergency area, see the receiving nurse to describe the symptoms and be registered. • In cases where the receiving nurse could quite accurately diagnose the problem right away, the patient would be sent directly to the next logical station, for example, X-ray, blood testing, a specialist, and so on. A nurse accompanied these patients to the next station.

154  Chapter Nine

• If the receiving nurse could not accurately diagnose the problem, the patient was given a number and placed in the waiting area, which was the case for the majority of the patients. • When called, the patient would see a receiving doctor, assisted by a nurse, where a first examination was performed to determine where the patient should go next. In some cases—for example, stitching small wounds—treatment could be given here and the patient might be allowed to go home. • Normally, the patient would be released back into the waiting area to await her or his place in line for the next step in the treatment, be it X-ray, blood testing, or another doctor. • When the patient’s number came up again, a nurse would take him or her to the next post. If this was only further testing—for example, an X-ray—the patient would be taken back to the waiting area after-­ ward to reenter the line. • Having received the test results, or if testing was not required, the patient would be taken to a specialist when the patient’s number in the queue came up. • For most patients this was the last step, either being treated and sent home or being diagnosed and having to come back during the day for further treatment or analysis. Statistics showed that many patients spent between three and four hours in the hospital before going home, many of them with only a cold, the flu, minor injuries, or other ailments that were quickly diagnosed and treated (sometimes simply with the order of going to bed). Personnel in the emer-­ gency ward complained that capacity problems in X-ray or other test facili-­ ties caused the long lines, while the specialists claimed they were ready and waiting for patients but sat idle for long periods of time before any showed up. To understand what was causing the problems, a bottleneck analysis was performed. The result is shown in Figure 9.28, which clearly indicates that the bottleneck was the admitting nurse. After further discussions, it became clear that the reception station was severely understaffed, and nurses were prevented from asking sufficient questions of the patients to allow them to be sent to the right place from the beginning. Adding one or two nurses here would change this and would give them the opportunity to do the first diagnosis here.

Analysis tools are quite different in nature and in what angle they take when viewing the data. This means that some of them lend themselves more naturally to one type of issue, and others lend themselves to other performance shortcomings. I suggest you consider this carefully before choosing a tool.

Waiting line

Receiving nurse

Receiving doctor

Waiting line

Blood test

Waiting line

Specialist

Cash desk

Figure 9.28  Emergency ward bottleneck analysis.

Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity Capacity available demand available demand available demand available demand available demand available demand available demand available demand

Receiving nurse

i

X-ray

Capacity Capacity available demand

Tools for Analyzing the Performance Shortcoming  155

10 Tools for Generating Ideas and Choosing among Them

T

he two preceding chapters described tools for two ­somewhat-­related stages of collecting data about the performance shortcoming and analyzing the shortcoming and its causes. Knowledge about the situation and its causes can be used at the next stage to generate ideas about possible solutions or improve‑ ments. This chapter describes some tools that can be used to generate ideas and aid in reaching consensus solutions: • Brainstorming • Brainwriting/Crawford slip method • Six thinking hats • Nominal group technique • Paired comparisons I have already mentioned that some tools and techniques can be applied elsewhere in the improvement process than where they were presented. This is especially true for these tools, as there will typically be many times during a project when there is a need to come up with a broad range of ideas and choose among them.

10.1  Brainstorming Finding solutions to the problems we are usually faced with in improvement proj‑ ects requires creativity. Most of us are far better at thinking analytically to find the right solution. In improvement work, the purpose is to find as many solutions as possible and then proceed with only the most promising ones. For this purpose, brainstorming and other related techniques described in this chapter are i­nvaluable.

157

158  Chapter Ten

Brainstorming is perhaps the ­best-­known approach, with the exact purpose of cre‑ ating as many ideas as possible, and it has numerous benefits: • Stimulates creativity • Encourages joint problem solving • Makes it possible for participants to build on one another’s ideas • Minimizes the tendency to prematurely evaluate ideas • Appreciates thinking that expands the traditional boundaries of the solution space There are at least two different ways to brainstorm: • Structured brainstorming, where each participant in turn launches one idea. This approach is more structured and ensures equal participation, but it is less spontaneous and to some extent limits the possibility of building on one another’s ideas. • Unstructured brainstorming, where everyone can freely launch ideas all the time. This approach is very spontaneous, but it is often more confusing and can lead to one person or a few people dominating the activity. For both ways, the procedure is as follows:

1. Clearly define the topic of the brainstorming and write it at the top of a white board or flip chart.



2. Let the participants launch ideas according to the approach used, struc‑ tured or unstructured. Encourage everyone to launch ideas, no matter how silly they might seem.



3. Write down every idea launched, preferably using the same wording as the original proposition.



4. Do not allow participants to discuss, criticize, or evaluate ideas during brainstorming. If an idea is not lucid enough to give any meaning, further explanation can be solicited.



5. Allow one break period when the flow of ideas stagnates; it will usually pick up again. When the ideas seem to become reformulations of previ‑ ously launched ideas or when the frequency of new ideas is decreasing for the second time, close the process.



6. At the end, evaluate the ideas. Start by picking the obviously good ideas, the “stars.” The rest of the ideas can be sorted into groups, either by theme or by decreasing potential.

Tools for Generating Ideas and Choosing among Them  159

The consolidated list of ideas will now form a suitable starting point for choosing the improvement ideas to proceed with. Also remember the established rules for brainstorming: • No criticism or discussion of ideas is permitted during brainstorming. • Laugh with crazy ideas, not at them. • Be loose and spontaneous; there are no stupid ideas. • Keep all ideas. • Combine ideas. The ideas are not mine or yours; they are the group’s.

Example Employees of a day care who worked with a class of 30 children, aged one to four years, received complaints from parents that their kids talked about episodes of teasing and bullying. The employees had also noticed that some kids made fun of others, but they did not consider it serious. Learn-­ ing that some children reacted strongly, the day care brought in an expert on antibullying. She suggested that the children be taught how to brainstorm ideas. The children were enthusiastic and participated in the training. She used examples such as games to play at birthday parties and healthful snacks and meals. Then she asked the children to brainstorm ways to achieve a good atmosphere in the classroom and avoid making anyone feeling sad. The children came up with more than 40 ideas, many of them very good; for example: • Create buddy pairs of one older and one younger child, and let this rotate every week • Take turns deciding which games to play, indoors and outdoors • No dessert at lunchtime for anyone caught teasing or bullying Altogether, several of the ideas were implemented and seemed to create a much better climate and play environment for all the kids.

10.2  Brainwriting/Crawford Slip Method Brainwriting and the Crawford slip method are two methods of brainstorming in writ‑ ing. The advantage of generating ideas through writing is that it is easier to describe more detailed and coherent ideas, which often leads to the development of equally coherent solutions. While brainwriting is almost solely a written adaptation of brain‑ storming, the Crawford slip method is a technique that protects the anonymity of the

160  Chapter Ten

participants. This second approach is therefore often used when the group experiences conflict that stifles creativity or when large amounts of information are expected to surface through the idea-generating session (Swanson, 1995). Looking at brainwriting first, it can be performed in two different ways: • The card method, where ideas are written on small cards and circulated among the participants for the addition of related ideas or expansion with other elements. • The gallery method, where ideas are written on a number of white boards or flip charts and the participants circulate among these to add related ideas or expand on existing ones. The procedure for brainwriting is as follows:

1. As with brainstorming, start by clearly defining the target topic for the idea generation. The topic can be written on either a white board or the partici‑ pants’ individual cards if the card method is used.



2. The participants then write down their ideas, either on the cards or on the white board. Precise formulations are encouraged so as to enable under‑ standing without explanation from the owner.



3. The participants are allowed to add to others’ ideas to reap the effects from combining ideas or further developing them.



4. At the end, the ideas are verbally discussed by the group and preferably also sorted into classes of ideas.

The Crawford slip method is a variant of the brainwriting card method in that the cards are not circulated for expanding on others’ ideas. Neither method is an open evaluation of the ideas conducted, which further increases the neces‑ sity of accurate idea formulation. Sorting of the ideas is done by one person, often aided by computer software designed for this purpose (several of which are offered on the market). The final document summarizes all ideas and can be openly used in the group to decide which ideas will be used in the improve‑ ment work.

Example A stable housed a riding school and a horse club, and private horse owners also kept their horses there. For years there were complaints and clashes between the different user groups. The horse club members were angry with

Tools for Generating Ideas and Choosing among Them  161

the riding school for continually occupying the riding hall and paddock, the horse owners claimed the riding school students borrowed or stole their equip-­ ment, and everyone blamed the others for not cleaning up after themselves. After the disappearance of an expensive saddle from the tack room used by the horse owners, the situation was close to erupting. The stable owner decided to do something, but he figured brainstorming would not be productive given the tension among the groups. He passed out suggestion cards to all users of the stable and asked them to write down as many pos-­ itive ideas as possible that could help make the place more harmonious. After receiving some 20 idea cards from all three user groups, he sat down to review and sort the ideas. Some had clearly not accepted his request for positive ideas and suggested he throw people out of the stable, close the riding school, and worse. Many ideas were genuinely construc-­ tive, though, and he ended up with quite a few suggestions he thought could be implemented: • Let the riding school and the riding club alternate between semes-­ ters in choosing paddock and riding hall hours • Seek permission from a neighboring farmer to use one of his fields as an alternative riding area • Take the riding school students into the woods for trekking now and then • Make sure the riding school equipment is always in order, to avoid having students look for replacement equipment elsewhere in the stables Although some people left the stable, the changes implemented improved the situation considerably.

10.3  Six Thinking Hats Six thinking hats is a method for actively encouraging people to view a problem and its solutions from several different perspectives. As such, it is considered more a technique for inspiring creativity than for generating ideas. However, I have come to view the technique as an ­idea-­generation method. Taking distinctly different angles in a discussion seems to bring up ideas that would otherwise not have surfaced. And deliberately having a team member take the position of devil’s advocate ensures that suggested solutions are debated with regard to feasibility and possible flaws. The technique achieves this by encouraging you to recognize what type of thinking you are using and to apply different types of thinking to the subject. Edward de Bono (1985) originally created the method when working to improve

162  Chapter Ten

creativity and lateral thinking. He gave the six thinking hats the following colors and respective roles: • The white hat: cold, neutral, and objective; the person “wearing it” should be systematic and careful in looking at the facts and figures • The red hat: represents anger; the wearer should listen to intuition/gut feel‑ ing and his or her own emotions • The black hat: pessimistic and negative; thinking with this hat on should focus on why an idea will fail • The yellow hat: optimistic, sunny, and positive; the wearer should focus on seeing ways ideas will work and overcoming obstacles • The green hat: represents grass, fertility, and growth; the person underneath it should be creative in cultivating new ideas • The blue hat: connected with the sky; the person should focus on seeing things from a higher perspective Although some of the connotations used in describing the hats can be construed as negative, it is important to understand that each hat is equally important in ensur‑ ing fruitful discussions. The steps in using the six thinking hats are the following:

1. Assign hats to the people in the discussion team, preferably one color per person, and make sure that everybody understands that when speaking during the session, he or she must clearly identify with the color of his or her hat



2. The team engages in a creative discussion about the problem, and each hat has the following responsibilities (other team members may contribute at each step): • The white hat presents facts about the problem • The green hat presents ideas on how the problem could be solved • The possible solutions are discussed, with the yellow hat focusing on benefits and the black hat on drawbacks • The red hat works to elicit all team members’ gut feelings about the solutions • The blue hat summarizes the discussion and closes the meeting

Tools for Generating Ideas and Choosing among Them  163

Example A movie theater with 12 theaters received many complaints from the clean-­ ing staff about the state of the theaters after shows: litter filled the rows between the seats, soda was spilled on the seats, and the backs of the seats were covered with dirty shoe marks. The manager assembled a team of six people—two members of the cleaning staff, two employees from the snack bar, a ticket salesperson, and himself. Having heard about the six thinking hats from a friend, he wanted to try this approach. He asked his friend, who had quite a bit of experience in using the method, to come teach them all about it. Spending half a day listening to this individual and discussing the approach, the team felt ready to proceed. They met one morning before normal working hours and assigned the hats among them. Proceeding with the discussion, they more or less managed to stay true to their parts, although there were some slippages. At the end, the manager, wearing the blue hat, summarized all the ideas that had surfaced. The most promising ideas were serving soda in bottles instead of paper cups, installing drink holders on the seats, placing better door mats at the entrances to the theaters, and taking up an old tradition of running ­commercial-­like messages about proper cinema behavior before the shows.

10.4  Nominal Group Technique As mentioned in the description of brainstorming, it’s possible that one per‑ son or a few people will dominate the activity. An unfortunate effect of this, besides the fact that the full potential of ideas is not realized, is that those who feel they have been overlooked will not commit to the produced conclu‑ sions later on. The intention of nominal group technique (NGT) is to render possible a brain‑ storming session where all participants have the same vote when selecting solu‑ tions. To use NGT, the following steps are recommended:

1. As in brainwriting, the first step is written idea production, where each person generates ideas and writes these on idea cards, one idea per card.



2. All the produced ideas are registered on a flip chart and the ideas are briefly discussed. The purpose is to clarify the content of each idea as well as eliminate similar ideas. At the end of this stage, each idea is assigned a letter, starting with “A.”

164  Chapter Ten



3. The next step is again an individual activity, where participants rank the ideas. From the complete list of ideas, each participant selects up to five ideas and writes these on his or her ranking card. Each idea is identified by its assigned letter from the list on the flip chart. When ranking ideas, participants assign points to the ideas—from 5 for the most important or best idea down to 1 for the least important or worst idea.



4. The session leader collects the ranking cards and writes the assigned points on the flip chart. For each idea, the points are summarized to total scores. The idea achieving the highest total score is the group’s prioritized idea or solution.

In the ensuing improvement work, start with the solution achieving the highest score or the two or three most important ones.

Example A company with about 400 employees experienced grave delays in the internal communication among individuals and departments. Normal brainstorming sessions produced some good ideas, but most people did not launch any ideas. NGT was therefore used next, and after the individu-­ ally generated ideas were collected, the list looked like this: A. Extend the use of e-mail B. Anyone who receives an inquiry should respond confirming they received the question C. Develop routines for keeping everyone informed when employees change offices D. Implement a system for telephone messages when traveling E. Have weekly department meetings and meetings across depart­ments F. Sort mail automatically G. Rebuild the office area into an open-plan office H. Have compulsory table rotation during lunch breaks The ranking card of one of the participants is shown in Figure 10.1, and the resulting flip chart list is shown in Figure 10.2. The result was that the company extended the e-mail system by giving everyone access to it and providing training in using it.

Tools for Generating Ideas and Choosing among Them  165

Problem: Poor internal communication Idea

Points

A

5

B C

4 1 2

D H

3

Figure 10.1  Individual ranking card. Problem: Poor internal communication Idea

Points

A

55453254

33

B C

434322 14535

18 18

D E

2345223 121

21 4

Total

F

23241

12

G

15321

12

H

35551

19

Figure 10.2  Resulting flip chart list.

10.5  Paired Comparisons A success factor in using NGT is that those involved manage to distinguish between the ideas and prioritize them. In real life, with many ideas to choose from, it can be difficult to decide which idea to vote for, and the results might be determined by coincidence. Like the other methods described in this chapter, paired comparisons aims at prioritization and consensus reaching, but it does so through sequentially comparing two items, instead of having to consider several items at once. Single decisions are easier to make than decisions that require you to select from a large number of possible solutions. The steps in paired comparisons are the following:

1. Clearly identify the alternatives to be compared. The total number, denoted N, should be manageable—that is, not more than eight.



2. Make a matrix with the alternatives, coded by letters, as row headings and the pairs as column headings, indicated by letters only, to save space. The num‑ ber of pairs p is determined by the following formula: p = [N × (N – 1)]/2.

166  Chapter Ten



3. Column by column, each participant votes for one of the alternatives; the votes are logged in the matrix.



4. After participants have voted for all pairs, add up the total number for each pair; this should equal the number of participants.



5. Add up the number of votes for each alternative to get the row totals. The ­highest-­scoring alternative is the preferred one, according to the group.

Example Imagine that the company from the NGT example wanted to apply paired comparisons instead. To avoid ending up with too large a matrix, only the five ­highest-­ranked ideas from the NGT session are brought forward here: A. Extend the use of e-mail B. Anyone who receives an inquiry should respond confirming they received the question C. Develop routines for keeping everyone informed when employees change offices D. Implement a system for telephone messages when traveling H. Have compulsory table rotation during lunch breaks The resulting comparison matrix is shown in Figure 10.3. You might notice that with this method, ideas A and D changed places. A/B

A/C

A/D

A/H

A

6

2

5

3

B

1 2

D

B/D 4

4 7

7

7

C/H

D/H

Total 10

1

7

H

C/D

5

3

5 7

B/H

16 0

C

Number of votes

B/C

6

5 2

3

7

7

7

2 7

7

7

4

10 23 11

Figure 10.3  Paired comparisons matrix for internal communication.

You should now be ready to unleash the most exciting and powerful tools: the improvement tools.

11 Tools for Creating Improvements

T

he categories of tools presented so far have all played important parts in the introductory phases of the business process improvement model. The main purpose of these tools is to contribute to improvement, even if they cover only a small portion of the total activity leading up to this target. The tools presented under the heading of improvement are at the outset more directly geared toward this purpose. The following tools are presented in this chapter: • Streamlining • Idealizing • QFD • Statistical process control/control chart • Six Sigma • Business process reengineering • Benchmarking

11.1  Streamlining The principle behind streamlining is to trim away excess waste and superfluous elements in business processes. A characteristic of a streamlined process is that it flows easily and without resistance that lowers the performance level within the process or within the environment in which it exists. Streamlining of busi‑ ness processes is really a collective term for several smaller techniques. These can be used completely on their own, but when put together, they produce a much stronger streamlining effect. Some of these techniques are briefly presented in the

167

168  Chapter Eleven

f­ ollowing sections (there are others, but either they are quite similar to those pre‑ sented or they are aimed at objectives already covered in previous chapters). They are presented in the order in which their use is recommended when streamlining a business process (Harrington, 1991). You will notice that streamlining becomes even more powerful when com‑ bined with the process model or flowchart you have made. Think of the flowchart as a map with a traveling route penciled in. When performing the streamlining exercise, you are identifying steep hills and deep valleys that are difficult to cross. These obstacles are marked with different colors in the flowchart, making it easy to see where the process flows slowly and progress is halted. Either these areas can be eliminated or the traveling route can be retraced to avoid them. 11.1.1  Bureaucracy Elimination The words streamlined and bureaucratic are quite opposite. Streamlined is usu‑ ally associated with a process that is efficient and flows easily; bureaucratic is synonymous with a process that is slow and cumbersome. Bureaucracy is often an obstacle on the way toward process thinking and the transition from departmental to process management. A natural first step when streamlining business processes, therefore, is to eliminate bureaucracy. One typical effect of bureaucracy is unnecessary paperwork. Many managers spend 40 to 50 percent of their time writing or reading ­job-­related material. Studies have shown that approximately 60 percent of all office work in a company consists of reviewing others’ work and storing and retrieving information—information that is sometimes useful but very often not. The negative effects of this are many and often impossible to measure. A critical review of all such bureaucratic tasks is therefore important for the clear purpose of minimizing all delays, red tape, and other operations that do not stand out as useful, as ­value-­adding, or as supporting other processes. Here is one possible approach to eliminating bureaucracy:

1. To start, the bureaucracy must be tracked down. Bureaucracy is often iden‑ tified by asking questions like: • Is the activity performed to inspect or approve someone else’s work? • Does it require more than one signature? • Are several copies made of the result? • Are several copies stored for no apparent reason? • Are copies sent to people who do not need or use them? • Are people or departments involved that stifle efficiency and the quality of the work?

Tools for Creating Improvements  169

• Is there unlimited use of the copier or access to large filing cabinets? Studies have shown that about 90 percent of all documents stored in an organization are never used again.

2. Activities that constitute bureaucracy should be colored blue in the flow‑ chart or process model for the process.



3. The person responsible for each blue activity presents an overview of time and costs related to the activity as well as its usefulness. This usually meets with some resistance, as all such activities are viewed by someone as essential to the daily operation and ­long-­term survival of the organiza‑ tion. One study documented a situation in which a request for the purchase of equipment exceeding a certain dollar amount had to go through five administrative levels for approval. Of 10 requests sent for review, two con‑ sisted entirely of blank pages behind the cover page allotting space for the signatures. Together with the other eight, these two requests passed right through the system and were duly approved.



4. Activities that cannot be justified are eliminated.

11.1.2  Redundancy Elimination Especially for administrative processes, it is true that many identical or similar activities are often performed at two or more places in the process. This occurs when different departments or different organizations in the supply chain perform their tasks independently and without knowing what the others do. This increases process costs and also increases the likelihood that conflicting data exist. An exam‑ ple is a case where the purchasing department solicits prices for a component. At the same time, the design department is also soliciting prices but is quoted a dif‑ ferent price. At different places in the organization, these data are used, causing inaccuracy and errors. Due to a lack of trust in many organizations, each person and each department often keeps their own records over matters like absenteeism and overtime. At the same time, this is done centrally, and the information in the various systems often differs. The costs this double record keeping entails and the problems caused by con‑ flicting data must be avoided. There really is no trick to this, beyond going through each of the activities and results associated with the process and identifying and eliminating activities, documents, and other results that are generated more than once. In the flowchart, these activities should be colored maroon. 11.1.3 ­ Value-­Added Analysis Value-added analysis is a central concept in the streamlining of processes. Let us first study the concepts of value and ­value-­added. When a product or service passes

170  Chapter Eleven

through the organization and is transformed from raw material into a finished prod‑ uct or service, two things happen to its value: • By supplying materials, labor, energy, and so on, the process incurs costs for the organization. The added value experienced by the product is, however, independent of these costs. • By adding functionality, aesthetics, brand name, and so on, to the product, it experiences an added value that makes it possible to sell it for a higher price (hopefully) than the collective costs the process has incurred. The challenge facing the organization is to ensure that the value of the product or service, expressed in terms of what the market is willing to pay for it, is higher than the costs of generating it. This way, value is actually a theoretical concept that expresses both market value and true material value. Added value (AV) can be expressed as follows:

AV = Va – Vb,

where Va is the value after processing and Vb is the value before processing. As already mentioned, value in this sense is affected by a number of more subjec‑ tive factors such as functionality for the intended use, prestige, synergy with other products or services, and so on. Furthermore, this value for the customer is inde‑ pendent of how much it costs to manufacture the product or deliver the service. If it were possible to manufacture a Mercedes for half the current cost, the value for the customer would still be the same. During the processing of the product or service within the organization, it is subject to a number of activities, which can be divided into three categories: • Real ­value-­adding activities (RVA)—activities that from the end customer’s point of view add value to the product. These are the typical transform‑ ing operations that create the product’s functionality and appearance or the service’s usefulness. • Organizational ­ value-­adding activities (OVA)—activities that from the customer’s point of view do not add any value, but that are necessary from the organization’s point of view. These can be production planning, mainte‑ nance, personnel administration, and so on. • Non-value-adding activities (NVA)—activities that from both the custom‑ er’s and the organization’s point of view add no value at all. Typical exam‑ ples are waiting, storing, reworking, and so on.

Tools for Creating Improvements  171

Value-added analysis looks at each activity in the process to determine its contribution to the added value for the end customer. The intention is to classify each activity as one of these three types and then work to minimize OVA and eliminate NVA. The analy‑ sis is carried out as shown by the sequence of questions in Figure 11.1 (Harrington, 1991). After each activity in the process is classified as one of these three categories, markers are used to color them in the flowchart. RVA are marked green, OVA yel‑ low, and NVA red. This gives a good overview of the portion of activities that are truly ­value-­adding. Such a picture can often shock the organization. Very often, less than 30 percent of the incurred costs is spent on RVA, while less than 5 percent of the time consumed belongs to RVA. Another way of portraying this information is by using a ­cost-­cycle time chart, as shown in Figure 11.2 (Harrington, 1991). The chart shows the effects that can be achieved by eliminating NVA and minimizing OVA in order to let RVA constitute the main part of the process’s time and costs. The core task of removing NVA and minimizing OVA is a project of its own, and there is no consensus as to how it should be performed. Some general advice includes the following: • Rework can be eliminated only by removing the source of the error • Pure movement of documents or other information can be minimized by combin‑ ing operations, moving people closer to each other, or automating the process Activity

Yes

Necessary to produce output?

No

Contributes to customer satisfaction? Yes

Contributes to the organization’s needs? Yes

Real valueadding

Organizational value-adding

• Order handling • Manufacturing • Packaging

• Planning • Maintenance • Hour registration

Activities that must be performed to satisfy customer requirements

No

No Non-valueadding • Storing • Inspection • Rework

Activities that do not contribute to satisfying customer requirements and that should be optimized or eliminated

Figure 11.1  ­Value-­added analysis questions.

172  Chapter Eleven

• Waiting time can be minimized by combining operations, balancing the work load, or automating processes • Most output from NVA can be eliminated if management accepts it • Inspection and control can be eliminated by changing policies and procedures The result should be an increase in the proportion of RVA, a reduction in the propor‑ tion of OVA, and a minimization of the proportion of NVA. After a typical ­value-­added analysis, the resulting chart might look like the one in Figure 11.3 (Harrington, 1991). 11.1.4  Process Cycle Time Reduction Critical business processes are subject to the rule that time is money. Such pro‑ cesses are usually carried out using critical resources that are often bottlenecks. At the same time, the products or services from these processes are the ones that really matter to the customers. Therefore, the products or services should be delivered to them as quickly as possible. Long cycles require capacity of the critical resources for unnecessarily lengthy periods of time, they stifle efficient delivery to the cus‑ tomer, and they incur storage costs for the company. To determine where the crusade against long cycle times should be started, consider activities that cause delays, activities that originally display a long cycle time, or particularly critical processes. Typical efforts to remedy long cycle times are the following:

Incurred costs

RVA OVA NVA

1

2

3

4

5

6

Cycle time

Figure 11.2  ­Cost-­cycle time chart.

7

8

9

10

Activity

Tools for Creating Improvements  173

Incurred costs Costs before RVA OVA NVA

Costs after

1

2

3

4

5

6

7

Activity

Cycle time after Cycle time before

Figure 11.3  Corresponding chart after completed ­value-­added analysis.

• Perform activities in parallel instead of in sequence. Very often, most of the steps in a business process are performed in sequence, although they could just as well be carried out in parallel. A serial approach results in the cycle time for the entire process being the sum of the individual steps plus trans‑ port and waiting time between steps. When using a parallel approach, the cycle time can be reduced by as much as 80 percent while producing a better result. A classic example is product development, where the trend for many years has been toward concurrent engineering. Instead of the sequence of a concept, then drawings, then bill of materials, and then processes, all activi‑ ties take place in parallel in integrated teams. The development time is dra‑ matically reduced, and the needs of all those involved are considered during the development process. • Change the sequence of activities. This point is closely connected to unnec‑ essary transport of documents and products. Often, documents and products are transported between machines, departments, buildings, and so on. For instance, a document might travel between two offices a number of times for inspection and signing. If the sequence of some of these activities can be changed, it might be possible to perform all processing of the object when it comes in the first time. • Reduce interruptions. Interruptions cause delays and increase the cycle time for critical business processes. The production of an important order can, for example, be stopped by an order from a far less important customer, but

174  Chapter Eleven

one that has turned into a rush order because it has been delayed. People working inside critical business processes can be interrupted by phone calls that could have been handled by someone else. The main principle is that everything should be done to allow uninterrupted operation of the critical business processes. • Improve timing. Many activities are of such a nature that they are performed with relatively large time intervals between each activity. For example, there might be a report that is generated only once a week or purchasing orders that are issued every other day. People using these reports should be aware of such deadlines to avoid missing them. Often, the manufacturing depart‑ ment does not know that purchasing requirements that are not reported by noon Thursday will not be purchased until the following week. Improved timing in these processes can save many days in cycle time. There are numerous other approaches suitable for reducing cycle times, but these are some of the most important ones. If you combine the streamlining analyses and consistently use color codes in the flowchart, the end diagram can become highly revealing. A general example is shown in Figure 11.4, portraying an imaginary process where steps representing bureaucracy, redundancy, NVA, and OVA are shown.

Redundancy

Bureaucracy

NVA

OVA

Figure 11.4  A flowchart with steps distinguished according to streamlining results.

Tools for Creating Improvements  175

Example After 25 years of continuous growth, a manufacturer of electric appli-­ ances had come to a point where many of the administrative business processes were cumbersome. Some important processes, especially order processing, had become so time-consuming that they inhibited further improvements in other operative processes. The company thus decided to streamline this process. As the company grew, a number of different computer systems had been installed, including an order system. However, when this system was installed, the old routines had simply been automated or transferred to an electronic format. These routines were designed for a much lower sales volume and complexity. The dominant opinion was that a large bureau-­ cracy was being kept alive artificially, and that the first step was to identify and eliminate the elements of this bureaucracy. The company was particularly interested in procedures and mecha-­ nisms where documents and decisions needed approval from one or more people or departments. An assessment of the system for order process-­ ing of production identified 12 instances where somebody had to approve someone else’s work. An analysis of these 12 instances showed that they consumed nine days altogether. At the same time, only two of these were true reviews undertaken before approval was given. Combined with fre-­ quent customer complaints about long ­ order-­processing times, it was obvious that something needed to be done. In the 10 instances where no real review was performed, the operative people performing the original work were given authority to approve their own work. This saved seven to eight days on average and an unknown amount of money. Furthermore, they knew that many subsystems had been connected throughout the years. Each of them performed the same or similar tasks. The logical step, therefore, was to perform a redundancy elimination. For this task, a detailed flowchart portraying the process and output from it was used. Upon closer inspection, 16 highly similar steps were found—mainly activities where different versions of order documents were generated. If one globally available order document was produced, 13 steps in the pro-­ cess could be removed immediately. Over a period of four months, the order system was totally redesigned to allow information that was entered once to be available to the entire organization. This eliminated nine days of work and dramatically improved the quality and accuracy of the process and its output. Finally, a ­value-­added analysis was undertaken. The current process contained a very small proportion of NVA. The proportion of OVA was a little higher, but not unreasonably so. By cutting the time consumed by individual activities related to registering internal documents, this portion could be further reduced. Figures 11.5 and 11.6, respectively, show the ­cost-­cycle time charts for the original process and the streamlined pro-­ cess. The cycle time was reduced by about 19 days, or 64 percent, and the costs by about $1000 for a standard order.

176  Chapter Eleven

Incurred costs

$2100

RVA OVA NVA

Cycle time 1 2

3 4 5 6 7 8 9 10 11 12131415 16 17 18 19 20 21 22 23 24 25 26 27 28293031 32 29 days

Figure 11.5  ­Cost-­cycle time chart for the original process.

Incurred costs $1100 RVA OVA NVA

Cycle time 1

2

3

4

5

6

7

Figure 11.6  ­Cost-­cycle time chart for the streamlined process.

8

9

10

9 days

11

Tools for Creating Improvements  177

11.2  Idealizing Instead of starting from the existing process and analyzing its individual steps for potential efficiency gains, as in the ­value-­added analysis, it is possible to free yourself from the limitations the current process imposes. This is the principle underlying ideal‑ izing. The purpose of this technique is to imagine how good the ideal process, without any form of waste or other inhibiting elements, could be. As a tool for generating pos‑ sible solutions, this can at times be quite effective. Even if it is obvious that the ideal process cannot be implemented in practice, it can provide insight into how it should be implemented. The differences between the ideal process and the current situation can thus be used as a starting point for formulating solutions and improvement projects. Idealizing is typically done as a group exercise, to ensure that as many ideas as possible are captured. It is important that those performing the process be part of this group, as they are the ones who dream of the ideal process. At the same time, people unfamiliar with practical limitations can present refreshing ideas. Thus, a mixture should be included in the group. There really are no specific guidelines on how to use or perform idealizing. Flowcharts can be used in this type of exercise. After constructing a flowchart for the ideal process, it is interesting to compare it with the chart for the current pro‑ cess. The powerful graphic representation achieved by this approach is also a good starting point. By studying the deviations between them, it is possible to define measures to move the process toward the ideal state, as shown in Figure 11.7. Idealizing is closely related to both clean sheet business process reengineering (presented in section 11.6) and the A∆T analysis, whose objective is slightly dif‑ ferent (described in section 12.1).

A

A

B C

B D E

Current process

Ideal process

E

H G

F

H I

I J

K

Figure 11.7  Analysis of the current and ideal processes in idealizing.

L

178  Chapter Eleven

Example A ­ medium-­sized printing company that did a number of smaller jobs for different customers—typically advertising brochures, catalogs, and similar material—was troubled for a long time by errors in the finished printed material. The company decided to use idealizing to outline the ideal pro-­ cess that would ensure discovery of errors before actual printing. Seven people from different functional areas in the company took on this job and had many short meetings during a period of two weeks. The resulting ideal process, as well as the current process, is shown in Figure 11.8. The company spent the next three months moving toward this ideal process. Although quite a lot remained to be done before the ideal process was reached, not in the least due to the need for investments, the amount of errors discovered after printing was reduced from an average of 18 per month to 1 potential error. Ideal process

Current process

Customer

Customer

Jobs in digital format

Jobs on paper or film

Test print in color

Copying

Proofreading and customer approval

Test print in black and white

Printing and finishing

Plate production

Proofreading by customer

Complete printed material

Printing and finishing

Test run

Complete printed material

Control of test run against test print

Figure 11.8  Idealizing in a printing company.

Jobs in digital format

Plate production

Tools for Creating Improvements  179

11.3  QFD QFD was presented in section 5.6 as one approach that can be used in creating your business process improvement road map. As I have mentioned, QFD was originally used as a tool to facilitate ­customer-­oriented product development. Over time, the core of the technique has proved powerful enough to warrant application in many different areas. For our purposes, the tool works both for initial improve‑ ment planning and for more detailed process improvement analysis. The core of QFD is that it enables you to see relationships and connections between ends and means as well as how each alternative mean impacts the col‑ lective set of ends. This principle, performed through constructing the house of quality, is most suitable for identifying possible improvements in a business pro‑ cess—and not only improvements, but those most likely to produce the overall best effect for the requirements that process faces. The general house of quality was presented earlier but is shown again in Figure 11.9. Used for business process improvement, the house has several different fields: • What expresses the requirement set for the business process. The most important requirements are those that represent external customers, but also included are requirements from external suppliers and internal customers and suppliers, as well as other concerns internal to the organization. • To distinguish between the importance of each of these requirements, the field of Importance can be used to assign weight factors. • How shows the means used to satisfy the different requirements. If one requirement is a short operation time for a service process, the mean could, for instance, be expanding capacity in periods around demand peaks. • In the Relational matrix in the middle of the chart, the relations between ends (or requirements in this case) and means are analyzed. • In the roof of the chart, the How vs. How field, connections between the different means can be analyzed. If another mean for the service process is staff reduction, motivated by the company’s demands for reduced costs, the conflict between these two can be easily visualized in the roof. • In the same way as for products in the original use of QFD, the Why field can be used during process improvement to perform a simple benchmarking against other organizations’ processes. • Finally, the How much field is used to illustrate the results of the analysis. For each connection in the relations diagram, the importance factor for the requirements is multiplied by the weight factor for the relation, and the sum of these products for each requirement is placed in the How much field. The higher the score in this field, the more requirements can be sat‑ isfied by this mean.

180  Chapter Eleven

What

Importance

How vs. how How

Why

Relational matrix

Points

How much

Figure 11.9  The general house of quality, the chart used by QFD.

As for the application of QFD in improvement planning, this is best explained using an example.

Example Figure 11.10 presents a house of quality for the business process of product distribution. The improvement team working on this process had identified the customer requirements for the process and discussed which features of the process influenced these requirements the most. Unable to agree on this issue, the team members eventually decided to use QFD, resulting in the example diagram. The requirements to the process have been split into those posed by the customers and internal demands. Each of the requirements has been assigned an importance factor. For example, speedy and safe delivery are considered essential, while the internal wishes to keep overtime down and maintain a good inventory overview are ranked quite low. The typical situa-­ tion, which is completely in line with the norm, is that the external require-­ ments count more than the internal ones. The next step involved brainstorming a number of means or process features of the distribution process that were assumed to have a positive impact on the requirements. At each connection point between a demand and a process feature, an assessment was made as to what extent the mean impacts the requirement, which is denoted by the symbols used in the relations matrix. In addition, data for these requirements were collected or estimated for two competitors of the company. This information is presented in the benchmarking column on the ­right-­hand side. As can be seen, the com-­ pany trails with regard to safe delivery, but both competitors, particularly Company A, scored very well. Targets based on input from the competi-­ tors’ performance levels and strategic decisions have also been defined.

5

Low costs

6

No need for overtime

3

Low need for space

5

Good overview

3

Roof

Weight

Strong Medium Weak

9 3 1

3 4% 67

3 6% 90

3 13% 194 64 7%

2 9% 141

Figure 11.10  House of quality for process improvement.

Own company

Company A

Company B

Target

Close cooperation with customers

Computer-based ordering system

Careful handling

Low finished-good inventories

Good communication with manufacturing

Roof Strong positive Positive Negative Strong negative

70

Organizational difficulty Absolute importance adjusted for difficulty Relative importance adjusted for difficulty

8%

Relative importance

11% 102 1 7% 102

Absolute importance

1 8% 123

Target

82

Company B

9%

Company A

12% 105 3 22% 315

5 4 3 2 1

Own company

32% 281 1 19% 281

Benchmarking of metrics

Targets

22

4

No extra costs

2%

Correct shipment papers

Local warehouses

9

Sound work planning

8

Complete deliveries

30

No breakage

3%

10

13% 115 1 8% 115

10

Benchmarking of the process

Max. 150,000

Speedy delivery Safe delivery

Less than 1 repair per 60,000 miles

Internal demands

Customer demands

Customer and internal requirements to the process

Dependable trucks

Importance

Process features

Sound distribution planning

Tools for Creating Improvements  181

1

2

3

4

182  Chapter Eleven

For each mean, the team evaluated whether any relevant quantifiable targets could be defined. As the matrix reveals, a target was defined for the number of repairs beyond ordinary maintenance for the distribution trucks, that is, less than one per 60,000 miles. However, far from every mean has had targets defined this way. Therefore, in the field below, an assessment on an objective scale has been made of the company’s own performance level, that of the competitors, and the company’s targets. For dependability of the trucks, for instance, the company is far behind the competitors, while the target of one repair per 60,000 miles is quite ambitious. The correlation symbols in the relations matrix, or more correctly the weight factors for the symbols, are multiplied by the factors that represent the importance of each of the requirement elements. For each process feature, these products are added up and placed in the field of absolute importance. The numbers in this field show the impact that the improvement will have on the collective requirements. In the field immediately below, relative importance has been calculated, that is, how large a percentage of the total impact the single process feature accounts for. From these calculations, we see that the three process features that stand out are sound distribution planning, good communication with manufacturing, and a ­computer-­based order system. Furthermore, a factor has been estimated that indicates how difficult it would be to implement changes in these issues in the organization. The higher the factor, the more difficulties are expected to be encountered. The importance values are divided by these factors, which give the correspond-­ ing importance figures, adjusted for organizational difficulty. If this aspect is included, it seems as if sound distribution planning is the process feature that has the highest ability to impact the requirements and the easiest implemen-­ tation. This gives the company a clear sense of what element in the distribu-­ tion process to improve. It probably also makes sense to include the issue that achieved the ­second-­highest score. Finally, an evaluation has been performed in the roof of the house of quality with regard to synergy between the different means. If focusing on the distribution planning, we see that positive effects can also be expected from improved communication with manufacturing and a ­computer-­based order system. Local warehouses can, on the other hand, complicate the dis-­ tribution planning task.

Even if this example does not capture all aspects of the use of QFD for improve‑ ment, it at least demonstrates how the volume of the voice of the customer can be sufficiently turned up to be heard when creating process improvements.

11.4  Statistical Process Control/Control Chart A useful characteristic trait of this world and the processes carried out in it is that stabilized processes display a logical statistical behavior. This means that the

Tools for Creating Improvements  183

results from a process normally fall within certain specifiable limits. This fact can be used for monitoring and improving an organization’s processes. 11.4.1  Definitions of Variation Before the actual tool of statistical process control (SPC) is presented, it is neces‑ sary to define some central terms in statistics. First, it is important to understand the difference between chronic and sporadic variation: • Chronic variation is variation that is inherent in the process and is caused by a long line of factors, without one single cause for the deviation. It is, in other words, caused by something in the system; therefore, it is often referred to as common cause variation. This variation is normal for the process and must be expected and lived with, unless the process itself is changed. If, for example, the time it takes to complete a mail route in a large office building is measured, it will invariably vary due to factors such as the pace of the mail carrier, the number of letters to be delivered, the number and duration of breaks, the waiting time for elevators, and so on. • Sporadic variation, also known as special cause variation, is caused by fac‑ tors that appear infrequently and is often larger than the chronic variation. The deviation can usually be traced to one single cause and can be elimi‑ nated by removing that cause. For the mail route, this could be variation caused by a new mail carrier, incorrectly sorted mail, or an ­ out-­of-order elevator. The purpose of SPC is to classify variation according to these two groups and thus facilitate further improvement of the process. 11.4.2  Basic Statistics To understand the principles of SPC, it is absolutely essential to first understand the basic principles of statistics, on which SPC is based. Most processes studied in an organization can be described reasonably accurately using two variables: • The arithmetic mean value of the process measure being used—for exam‑ ple, the diameter of a hole being drilled. The statistical term for this value is the expected value, and the expression for it is:

X=

1 n ∑X . n i=1 i

• The standard deviation of the process—that is, a variable that says some‑ thing about how much variation should be expected from the process. If we want to drill a hole with a diameter of 7.9 millimeters—that is, an expected

184  Chapter Eleven

value of 7.9—the standard deviation could be 0.15 millimeters. The symbol for the standard deviation is σx, and the formula for calculating it is: n



σx =

∑(X i =1

n

− X )2

( n − 1)

.

Actually, it is not totally correct to use σ for the standard deviation. σ refers to the standard deviation for the distribution itself, while the standard deviation for the sample being taken is labeled s. This connotation has been used throughout the rest of the book. If the observation or measurements of the process are evenly distributed around the expected value, we can use a normal distribution to describe the process. This is often the case for stable processes. The normal distribution in statistical matters is a special distribution for several reasons: • It is symmetric; that is, the likelihood of the process giving a result that is larger than the expected value is as high as for a result smaller than the expected value. • If the distribution function graph is drawn, it looks like a bell; thus, another term for this distribution is bell-shaped. Figure 11.11 shows a distribution curve that expresses the likelihood that the process measure will assume a certain value. • From the graph, it can be seen that the distribution curve crosses the x axis at two points, these points being respectively three times the standard devia‑ tion to the left of the expected value and equally far to the right of it. This means that the probability of getting a value that is less than the expected value minus three times the standard deviation, or a value that is larger than the expected value plus three times the standard deviation, is practically zero. This is a basic quality of the normal distribution, namely, that the maximum deviation that can be experienced lies within 3σx. To be totally correct, 68.26 percent of all values will fall within one standard deviation, 95.45 percent within two, and 99.73 percent within three. The area below the distribution curve equals 1, which is another way of saying that all the probability is gathered within these limits of +3σx. Because all results from a process following this distribution are expected to fall within the total spread of six standard deviations, data that fall outside this area are sporadic deviations due to special causes. To monitor the process, control charts are used that show the process limits of three standard deviations in each direction (see Figure 11.11), as well as the continuous measurements of the process. The construction of such charts is discussed in section 11.4.4 on page 189.

Tools for Creating Improvements  185

Lower process limit, LPL x

Upper process limit, UPL x

Probability

0,5 A=1

3¦x

X

1sx

2sx

3sx

Figure 11.11  Normal distribution curve.

When it comes to the quality of a process, the narrower the chronic variation, the better the process. SPC is used to reduce this natural variation and to report occurrences of any sporadic variation, whose cause must be found. Before this is possible, the process must be statistically stable, that is, having a constant expected value and variation width. With these simple statistical principles and related rules, it is thus possible to control a business process and discover whether the centering is changing, whether the variation width is increasing, whether variation is seen beyond the expected limits, but also what normal variation must be expected. If a process measure lies far from the expected value but within three times the standard deviation, it might be tempting to adjust it. But this introduces lopsidedness into the process, and in the future, deviations to the other side will occur. Attempts to rectify this by adjust‑ ing the process the other way usually moves the deviation to the other side. This evil circle that can result from adjusting processes that display only normal varia‑ tion is popularly termed tampering and can be avoided by proper use of SPC. Far from just physical, geometric measures related to manufacturing processes can be kept under statistic monitoring. By using a number of different control charts, many different processes and process measures can be made subject to SPC: • The time it takes to perform a task, physical or administrative • The costs incurred by a process, physical or administrative • The number of errors or defects produced by a process, physical or administrative • Specific geometric measures such as length, diameter, and so on

186  Chapter Eleven

By defining control limits on the basis of the expected and normal variation inher‑ ent in the process, continuous measures of the process’s characteristic features are plotted in the chart. Depending on the results, different measures are taken. Which measures are appropriate in which situations is discussed in detail in sec‑ tion 11.4.5. Thus, SPC presents a powerful tool for monitoring and improving business processes. 11.4.3  Types of Control Charts The normal distribution and how control limits can be based on three standard deviations is the simplest type of control chart. This type is not suited for all sit‑ uations, however, so other types have been defined. First, there is a difference between two basic types of variables: • Variable data are variables that are based on measurement—for example, meters, hours, volts, and so on. These variables are measured on a continu‑ ous scale and with reasonably high accuracy. • Attribute data are variables measured by counting or classifying nonmea‑ surable characteristics (for example, acceptable/nonacceptable, brown/red, assembled/not assembled, and so on). Attribute data are quantified by inte‑ gers or intervals of values. Variable data can be converted to attribute data (for example, length by defining a limit for acceptable and nonacceptable). For variable data, a pair of charts must be used: one to control the centering of the process and one to control the variation width. For attribute data, only one chart is required. There are two weaknesses related to the use of a chart like the one shown in Figure 11.11: • The values that have been measured are not necessarily normally distrib‑ uted, which renders the process limits uncertain • Such a chart is not very sensitive, as values within the process limits can be measured long after the process has changed its center or the variation width has increased These weaknesses are usually remedied by using mean values for a group of mea‑ sures instead of single values as the basis for the chart. The advantage of this is that mean values for a group of measures, almost regardless of the original distribution of the individual measures, approximately follow a normal distribution. Thus, the first weakness is avoided. At the same time, a chart based on such values is far more sensitive to changes in the centering of the process, which also counters the second weakness. In opposition to the process limits in charts based on single values, the limits in charts based on average values are termed control limits. The

Tools for Creating Improvements  187

expected value for a mean variable consisting of the average m groups of single measures is then labeled:

X=

1 m ∑ X m. m j =1

The lower control limit (LCL) and the upper control limit (UCL), respectively, are calculated by subtracting and adding three times the standard deviation for the mean distribution. The standard deviation for a mean variable from a group of n individual values is calculated as follows:

sx =

sx n

.

There are seven chart types that are most frequently used, classified according to the variable being measured and the number of measurements that exist. These are shown in Figure 11.12 and described in more detail after a brief overview (Swanson, 1995). For variable data, there are three types of control charts: – – • X and s, where the X chart plots the mean value for groups of measure‑ ments, and the s chart monitors the variation width of the average measure‑ ments. The number of measurements in each group should normally be higher than 10. – – • X and R, where the X chart displays the same average measurements for the expected value. Because the number of measurements in each group is often lower than 10, and because the term range R, defined as the dif‑ ference between the largest and smallest value in the group, is easier to understand than s, an R chart is often used to monitor the variation width. – For this reason, the chart pair X and s will not be described further. • x and R, where the x chart is based on the type of individual measure‑ ments first described. This type of chart does not quickly detect changes in the process. The reason for occasionally using this chart is that the measurement costs are lower. At the same time, it may be the only alter‑ native if the variable changes very quickly, which makes it impossible to make group measurements—for example, during a measurement of temperature. The variation width is controlled using an R chart, but in this case, R means the difference between the current and previous measurement. Before the charts of attribute data are presented, it is important to understand the difference between defective and defects. Defective items are scrapped because they do not satisfy the requirements, and defects are units that have some kind of

188  Chapter Eleven

Chart type

Variable data

Control charts Defective Attribute data Defects

n > 10

X and s

n < 10

X and R

n=1

x and R

n is constant

p or np

n is variable

p

n is constant

c or u

n is variable

u

Figure 11.12  Types of control charts and when to use them.

weakness but can be accepted because the specifications allow some degree of weakness. For attribute data, there are four chart types: • np, which is used to monitor the number of defective items in a sample of constant size. The process measure used could be the number of defective cars manufactured in a day or the number of errors detected during order processing in a day. • p, which is used for the number of defective items in samples of either con‑ stant or varying size. • c, which is used to monitor the number of defects occurring in samples of constant size. Examples are the number of errors in a ­five-­page document or the number of defects found on a circuit board. • u, a more general variant of the c chart, which can also be used for samples of varying size. Notice that the np chart plots values that cannot exceed the sample size, because only a portion of the items can be defective. Because the c and u charts are used for the number of defects, this number might very well exceed the sample size. The differences between defective and defects can be further illustrated by an example. When the surfaces of 10 items were inspected, 5 items were found to have a total of 11 inconsistencies, while the other 5 were perfect. An np chart would plot five defectives if the specifications did not allow such inconsistencies. A c chart would plot 11, as this was the total number of errors, but would not say anything about how many would have to be scrapped.

Tools for Creating Improvements  189

11.4.4  Constructing Control Charts In this section, a general procedure for using SPC will be introduced. Afterward, specific instructions will be given for calculating control limits for the different types of charts and for how the charts are used. Generally, when using SPC, the steps are:

1. Determine which type of data will be collected and controlled—variable or attribute data—and the expected number of samples.



2. Following the guidelines of Figure 11.12, select a suitable type of control chart.



3. Collect the data required to calculate the control limits.



4. Construct the control chart by letting the y axis represent the variable being measured. The x axis represents the process and is divided into measurement segments (time, subgroups, measurement number, and so on). Mark suit‑ able intervals on the axes that make it easy to plot the measured values.



5. Calculate the lower and upper control limits and, where required, the mean and range of the data set. If one of the points falls outside the control lim‑ its, this point is removed and new limits are calculated. If further points fall outside, the process is not statistically stable and the cause must be found and eliminated before the use of SPC can be continued.



6. Draw the control limits in the chart along with the existing measurements.



7. Continue plotting values as they are collected from the running process. The meaning of the points and the necessary actions are interpreted fol‑ lowing the guidelines given in the next section.

While this is a manual procedure for using SPC, many PC programs designed for quality improvement contain modules for constructing control charts. The rest of this section provides detailed guidelines for constructing the individual chart types. – X and R Charts – As mentioned, the X and R charts are a pair of charts in which the first focuses on the centering of the process and the other controls the variation width. To construct such a chart pair, the following must be done:

1. Collect the data; typically, more than 125 measurements of recent date are required.



2. Divide the data points into logical subgroups, where at least 25 subgroups are required to calculate the control limits. Subgroups should consist of

190  Chapter Eleven

data points that logically belong together; that is, they are measured under the same conditions, close to each other in time. Usually, there will be two to five data points in each subgroup, which defines the sample size n. – 3. For each subgroup, calculate the mean value X and the range R (that is, the difference between the largest and smallest value in the subgroup). For the mean value, add one decimal place to what is used in the measurements themselves. – 4. Calculate the mean value for the entire data set X by either averaging all the individual measurements or averaging the calculated means for the sub‑ groups. Here, add two decimal places to what is used in the original measure‑ – ments. This value will represent the center line of the X chart. – 5. Calculate the average range R by finding the mean of all the range values calculated in step 3. This value represents the center line of the R chart.







6. Calculate the control limits for both charts. To simplify the calculations, calculation factors have been developed that correspond to a deviation of +/– three standard deviations for sample sizes from 1 to 10 (see Table 11.1). For – larger sample sizes, an X and s chart should be used. Notice that for all sub‑ groups of six or fewer measurements, no LCL for the R chart is calculated. – For the X chart, the control limits are calculated as follows, using two more decimal places than in the data points:

Table 11.1  Calculation factors. Number of data points in the subgroup n

Factors for the – X chart, A2

Factors for the LCL of the R chart, D3

Factors for the UCL of the R chart, D4

 2

1.880

—

3.267

 3

1.023

—

2.575

 4

0.729

—

2.282

 5

0.577

—

2.115

 6

0.483

—

2.004

 7

0.419

0.076

1.924

 8

0.373

0.136

1.864

 9

0.337

0.184

1.816

10

0.308

0.223

1.777

Tools for Creating Improvements  191

LCL X = X − A2 R ,

UCL X = X + A2 R.

For the R chart, the corresponding formulae for the control limits are as fol‑ lows, using one more decimal place than in the individual measurements: LCL R = D3 R ,



UCL R = D4 R.



7. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example, when the measurement was performed, by whom, or sample size). Draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.



8. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

Examples of these types of charts are shown in Figures 11.13 and 11.14. x and R Charts The x and R charts are a pair of charts used for the centering and the variation, respec‑ tively. As previously mentioned, this is a less sensitive chart than the preceding one, but one that is useful where measurement is very expensive or difficult. The construc‑ – tion follows almost the same steps as the X and R charts, using only slightly different calculation formulae for the control limits. The steps are as follows:

13.00

X

11.00

UCL = 10.78

9.00 7.00 X = 6.12 5.00 3.00 LCL = 1.46

1.00 n=6

– Figure 11.13  Example of an X chart.

Subgroups

192  Chapter Eleven

R 12.0 UCL = 11.3 10.0 8.0 6.0 R = 5.7 4.0 2.0 0.0 Subgroups

Figure 11.14  Example of an R chart.









1. Collect the data; typically, more than 25 measurements of recent date are required. – 2. Calculate the mean value for the entire data set X . This value will represent the center line of the x chart. 3. Calculate the ranges R for the data set, where R is defined as the numerical distance between one data point and the preceding one. Thus, for the first data point, no R is calculated, resulting in the number of R values being one less than the total number of measurements. – 4. Calculate the average range R for the entire data set by finding the mean of all the range values calculated in step 3. This value represents the center line of the R chart. 5. Calculate the control limits for both charts. To simplify the calculations, calculation factors have been developed that correspond to a deviation of +/– three standard deviations. For the x chart, this factor is 2.66, which gives the following calculation formulae for the control limits of this chart: LCL x = X − 2.66 R , UCL x = X + 2.66 R.

For the R chart, we know that the sample size is always two, as the range is the distance between two measurements. Thus, we can use the factor n = 2 from Table 11.1. For the UCL, we see that the factor is 3.267, while we remember that for sample sizes of six or fewer, there is not a lower limit. The expression for the UCL is:

UCL R = 3.267 R.

Tools for Creating Improvements  193



6. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example, when the measurement was performed, by whom, or sample size). Fur‑ thermore, draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.



7. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

The R chart will look exactly like the example in Figure 11.14, and the x chart will be very similar to the example in Figure 11.13, but with a different variable name for the y axis. p Chart Unlike the charts described thus far, which are applied to variable data, the charts for attribute data are not used in pairs. Each chart for attribute data covers both the centering and the variation width. The procedure to construct a p chart, where p represents the portion of defective units, is as follows:

1. Collect the data set to be used for the construction of the chart and divide the measurements into subgroups sorted, for example, by date or series. Generally, there should be more than 25 subgroups of a sample size n, with 50 or more in each subgroup, and the average number of defectives np should be larger than three or four. The data set should come from a process that was recently run.



2. For each subgroup, the portion of defectives p is calculated simply by dividing the total number of defectives by the sample size n and multiply‑ ing by 100 to produce a percentage figure.



3. The center line of the chart is the average defective portion for the entire data set p–, which is arrived at by dividing the total number of defectives by the total number of units inspected.



4. The control limits are calculated using the following formulae: LCL p = p − 3

UCL p = p + 3

(

p 1− p

(

n

),

p 1− p n

).

If the numerical value for the LCL is negative, no lower limit can be cal‑ culated, so it is left out of the chart. Also notice that as the sample size n changes with time, the control limits also change. This means that for each subgroup of a different sample size, new control limits must be calculated.

194  Chapter Eleven

As this is very cumbersome, a simplification has been devised that gives accurate results. If all subgroups have a sample size that lies within +/– 20 percent of the average sample size, then in the preceding formulae, the n of the denominators can be replaced by the average sample size n–. For sub‑ groups of sample sizes outside this 20 percent limit, separate limits must be calculated.

5. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example, when the measurement was performed, by whom, or sample size). Draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.



6. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

All four chart types for attribute data look like the preceding examples but use a different variable for the center line. Thus, no separate examples have been prepared. np Chart Whereas the p chart can be used for samples of varying size, the np chart can be used only for samples of constant size. Otherwise, the procedure is quite similar to the one used for the p chart:

1. Collect the data set to be used for the construction of the chart and divide the measurements into subgroups. Generally, there should be more than 25 subgroups of a constant sample size n of 50 or more, with an average num‑ ber of defectives np that is larger than three or four. The data set should come from a process that was recently run.



2. The center line of the chart is the average defective portion per subgroup np, which is arrived at by dividing the total number of defective units by the number of subgroups.



3. The control limits are calculated using the following formulae:

(

)

(

)

LCL np = np − 3 np 1 − p ,

UCL np = np + 3 np 1 − p .

If the numerical value for the LCL is negative, no lower limit can be cal‑ culated and it is therefore left out of the chart.

4. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example,

Tools for Creating Improvements  195

when the measurement was performed, by whom, or sample size). Draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.

5. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

u Chart As previously mentioned, the main difference between p/np charts and u/c charts is that the former are used for defective units and the latter are used for the number of defects. Defects do not necessarily mean that the unit is defective. The u chart is used for constant and varying sample sizes and is constructed according to the following steps:

1. Collect the data points; typically, there should be more than 25 data pairs consisting of the number of units inspected n and the number of defects found c. The data set should be recent, and the number of inspected units should be more than 50 with an average number of defects per subgroup of more than one to three.



2. For each subgroup, the portion of defects u is calculated (that is, the num‑ ber of defects c is divided by the sample size n). Because the sample size varies, this must be done for each subgroup.



3. The center line of the chart is the average defect portion u–. This is found by dividing the total number of defects in the entire data set by the total number of inspected units.



4. The control limits are calculated using the following formulae: LCLu = u − 3

u , n

UCLu = u + 3

u . n



If the numerical value for the LCL is negative, no lower limit can be cal‑ culated, and it is left out of the chart. Notice that as the sample size n changes with time, the control limits also change. This means that for each subgroup of a different sample size, new control limits must be calculated. A similar simplification to the one used for the p chart can be used in this case. If all subgroups have a sample size that lies within +/– 20 percent of the average sample size, the n of the denominators can be replaced by the average sample size n– in the preceding formulae. For subgroups of sample sizes outside this 20 percent limit, separate limits must be calculated.

196  Chapter Eleven



5. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example, when the measurement was performed, by whom, or sample size). Draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.



6. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

c Chart For the c chart, the chart for portion of defects in samples of constant size, the procedure is quite similar to that used for the u chart:

1. Collect the data points and divide them into subgroups. Typically, there should be more than 25 subgroups of more than 50 recent measurements, with an average number of defects per subgroup of more than two to three.



2. The center line of the chart is the average number of defects c–. This is found by dividing the total number of defects in the entire data set by the number of subgroups.



3. The control limits are calculated using the following formulae: LCL c = c − 3 c ,



UCL c = c + 3 c .

If the numerical value for the LCL is negative, no lower limit can be cal‑ culated and it is therefore left out of the chart.

4. Draw the axes, adding the correct variable notations. Divide the axes into suitable intervals and indicate any other relevant information (for example, when the measurement was performed, by whom, or sample size). Fur‑ thermore, draw the center line and control limits, using a solid line for the center line and dotted lines for the control limits.



5. Plot the data points as new measurements are made. Interpretation and actions are explained in the next section.

11.4.5  Interpreting the Control Charts It is important to be able to select the right chart type for a given situation and to construct that chart. The chart is useful only when we are able to interpret the meaning of the data points and understand what actions should be taken. Most of the variation seen in a control chart is chronic variation, which is a natural part of the process. This variation requires no adjustments or tampering

Tools for Creating Improvements  197

with the process to keep it under control. However, it should be a constant objec‑ tive to improve the process, that is, improve the expected value and/or reduce the natural variation width. If the process for issuing an important document is stabi‑ lized at an expected value of 3 hours and 20 minutes and a standard deviation of 30 minutes, SPC can very well be used for registering that after some time, it has been improved to an average of 2 hours and a standard deviation of 15 minutes. Only a small portion of variation can be attributed to special causes—what we called sporadic variation, which requires special attention or action. Results that fall outside the control limits obviously constitute such variation, as do patterns of variation within the limits, which can signal the existence of special causes. In the following list, some examples of control charts displaying different types of varia‑ tion and the meaning of these are presented:

1. The natural progression of a process under control that displays only varia‑ tion that can be statistically expected is shown in the previous examples (see, for instance, Figure 11.13). Unless there is a wish to improve such a process, there is no need to do anything about it.



2. Figure 11.15 shows a control chart in which one point has fallen outside the UCL. It is statistically very improbable that this should happen. With limits based on three times the standard deviation, only two or three of 1000 will fall outside. There is, however, sufficient probability that one isolated point, a wild card, can exist without there being any special cause for it. It is not alarming as long as no more points appear outside the limits in a short time.



3. If several points fall outside the limits, as shown in Figure 11.16, they undoubtedly are the result of special causes. A search for the causes must then be started in order to eliminate them, thus preventing a recurrence.



4. A process is changing or unstable if two of three consecutive points on the same side of the center line are more than two standard ­deviations away

UCL

Center line

LCL

Figure 11.15  Control chart with one point outside the control limits.

198  Chapter Eleven

UCL

Center line

LCL

Figure 11.16  Chart with several points outside the control limits.

from it, or if four of five consecutive points on the same side of the center line are more than one standard deviation away from it. One example is shown in Figure 11.17. Such a change in level indicates a special cause— for example, a new operator has taken over, a new material has been intro‑ duced, or a new tool has been put to use. If the cause is found, new control limits must be calculated under the new conditions before the monitoring is continued. Incidentally, if this is due to a change in operators, this is a good opportunity for improving the process. The operator producing the best results can teach the others what he or she is doing—a form of bench‑ marking between operators.

5. Another sign indicating that the centering of the process has changed is if relatively many points fall on the same side of the center line. Initially, we would expect an equal number of points on either side of the center line. If series of seven, 10 of 11, 12 of 14, or 14 of 17 fall on the same side, this indicates systematic variation. The causes of the change must be found and eliminated. An example is shown in Figure 11.18.



6. Furthermore, we expect that the measurements would alternately increase and decrease from the previous measurement. A trend where six or more measure‑ UCL 2 standard deviations 1 standard deviation Center line

LCL

Figure 11.17  Process with change in level.

Tools for Creating Improvements  199

UCL

Center line

LCL

Figure 11.18  A series of 12 or 14 points on the same side of the center line.

ments show a constant increase or decrease, as shown in Figure 11.19, is a signal that the process is changing. Again, the causes must be found and eliminated.

7. While alternating measurements should be expected on both sides of the center line, it is statistically improbable that 14 or more points should systematically alternate from one side to the other. This is shown in Figure 11.20, and it is probably caused by tampering with the process, that is, continually changing the centering as a result of individual measurements.



8. The last alternative when it comes to the appearance of the control chart is shown in Figure 11.21. Here, it is possible to see clear cycles of the measure‑ ments, which is probably caused by rotation of operators, a process span‑ ning several shifts, temperature changes throughout the day, and so on.

When using SPC, we must combine a basic understanding of the process being monitored with knowledge about the appearance of the control charts. In this way, SPC is a powerful tool that can be used both for inspecting the process and for improving it. Signals from the chart that indicate something is wrong must be followed up quickly. It is quite common in companies using SPC that when the UCL

Center line

LCL

Figure 11.19  Process with trend.

200  Chapter Eleven

UCL

Center line

LCL

Figure 11.20  Process probably involving tampering. UCL

Center line

LCL

Figure 11.21  Chart displaying cycles.

operator is not immediately able to determine special causes, a ­ cross-­functional team is established to continue working on the case. Such groups were discussed in section 6.4. 11.4.6  Process Capability Because it is connected to SPC, it is logical to cover the concept of process capa‑ bility here. This concept is used to determine whether a process, given its natural variation, is capable of satisfying the specifications it is expected to produce within. A process can be under statistical control and still not be able to produce units or results that satisfy the specifications. When calculating the process capability, an index is used to compare the specifications with the control limits. Specifications normally define a nominal value, the desired result, plus or minus a value that determines how large a deviation from the desired value can be accepted without negatively impacting the performance of the process. Normally, ­two-­sided tolerance limits are specified (both an upper and lower limit, UTL and LTL), but for some processes, only one of the limits is specified. The less complex of the two capability indexes, Cp compares the difference between the tolerance limits with the natural variation of the process. If the dis‑

Tools for Creating Improvements  201

tance between the tolerance limits exceeds six times the standard deviation, the process can stay within the specification, provided it has been centered on the nominal value. The index is calculated as follows: Cp =



UTL − LTL . 6σ

The standard deviation σ for the process is rarely known and must therefore be estimated. This is done by calculating the average range, that is, the difference between the highest and lowest value within a subgroup, for all subgroups. This is – usually R. An estimate of the standard deviation is calculated using the following formula: R σ= . d2 The factor d2 has already been calculated and can be found in Table 11.2, listed according to sample size n. A general rule for this simple index is that it should be greater than 1.2. This ensures that the process variation is sufficiently small compared with the specifi‑ cations. The problem with this index, however, is that it does not capture the fact that the process can be unevenly centered compared with the nominal value. It is of little help if the variation is small compared with the tolerance limits, if the center of it is located close to one of the limits. Therefore, a less sensitive index has been introduced, Cpk. This index covers the variation of the process as well as the centering of it. It actually consists of two

Table 11.2  Factors for estimating the standard deviation. Sample size n

d2

 2

1.128

 3

1.693

 4

2.059

 5

2.326

 6

2.534

 7

2.704

 8

2.847

 9

2.970

10

3.078

202  Chapter Eleven

indexes: one lower, Cpl, and one upper, Cpu. The smallest of these is used as Cpk. They are calculated as follows: C pl =

X − LTL , 3σ

C pu =

UTL − X . 3σ



– X in this case means the average value for the process. Values below 1.0 mean that a 100 percent inspection must be performed. If any changes in the process’s centering can be accepted, Cpk must be greater than 1.33, and preferably 2.0 or even larger. The ideal is that the index is over 5. This index can be used to deter‑ mine whether the existing process is at all able to produce what is required, to set realistic tolerance limits suited to the process, and to negotiate targets for suppliers’ process capability.

Example A manufacturer of various types of special cables experienced high quan-­ tities of cables with defects. In fact, defects were so common that pro-­ ducing excess quantities to accommodate for this had become the norm. The defects happened in spite of the stringent inspection along the entire manufacturing line. Operators were responsible for undertaking visual con-­ trol of the cable they were working on. To investigate what caused the defects and possibly correct them, the company wanted to try SPC. One day’s production was viewed as a sam-­ ple. Since a cable with defects in practice is always defective and has to be scrapped, a chart type covering defectives was selected. A p chart based on data from the previous month was first constructed. The center line was calculated at an average defective portion of 11.5 percent, which gave the chart shown in Figure 11.22. As the chart shows, the process was actually under control, with an average portion of defectives of 11.5 percent. This was shocking news that revealed how large a portion of defectives the company generated. To learn why the process was performing so poorly, the data points were divided into other subgroups. Instead of viewing the data from one day’s production, data representing a week’s production from each individ-­ ual operator were collected. A new p chart, shown in Figure 11.23, was constructed. This chart quite clearly reveals that while the other operators gen-­ erated few defectives, operators 5 and 12 were at the other end of the scale. After these two operators were studied more closely, it was found that both had such poor eyesight that they did not discover the many

Tools for Creating Improvements  203

p

19.00 UCL

16.00 13.00

p = 11.5

10.00 LCL

7.00 4.00 1.00

Working days

1

5

10

15

20

Figure 11.22  p chart for one day’s production. p

19.00 UCL

16.00 13.00

p = 11.5

10.00 LCL

7.00 4.00 1.00

Operator number

1 2 3 4 5 6 7 8 9 10 11 12 13 14

Figure 11.23  Control chart divided into operators.

smaller defects that caused defective units. The company covered the cost of glasses for them, which quickly gave results. The average portion of defectives was continually reduced and after three months was down to 1.85 percent. This number was still too high, but it was significantly better than the starting point. After this, the differences between the operators were so small that common charts could be used for the entire production of a day.

204  Chapter Eleven

11.5  Six Sigma Having included SPC as an improvement tool, it is virtually impossible not to men‑ tion Six Sigma. Some will undoubtedly argue that Six Sigma should precede SPC, and I will not argue about that, but the fact is that they are highly related. Histori‑ cally, SPC is an antecedent of Six Sigma (Quinn, in Bertels, 2003), but whether it is correct to include Six Sigma as a tool is debatable. As you will see for both business process reengineering and benchmarking later in this chapter, Six Sigma is more of a meta-tool. More specific tools and techniques are included in Six Sigma, and the concept itself outlines more of a model for how these are used in a concerted effort. In addition, Six Sigma is as much a leadership model and a state of mind as it is an improvement approach. The reason that it is related to SPC should, however, be obvious. Six Sigma takes its name (indirectly) from the +/– 3 standard deviations explained pre‑ viously and shown in Figure 11.11. It is generally accepted that Motorola pioneered Six Sigma from the mid-1980s onward, out of a realization that mea‑ suring defect rates as parts per hundred was far from good enough in the emerg‑ ing global competition from Europe, Japan, Korea, and China. Staying within the +/– three sigma limits, 0.135 percent of the output is still unacceptable. To win that game, defects need to be measured in parts per million (ppm), which is how current purchasing agreements in many industries—for example, automo‑ tive and electronics—define quality levels. By reducing the variation (sigma) until the distance between the mean value and the specification limits is six sigma, much better performance is achieved. Even though the specification lim‑ its are not moved, since the value of sigma is smaller, the new average defect rate is only 3.4 per million. One of the first companies to win the newly introduced Malcolm Baldrige National Quality Award in 1988, Motorola shared its methods broadly, making its Six Sigma Quality program accessible to other companies. The number of com‑ panies using Six Sigma, along with the number of books and consultants offering advice, has dramatically grown since then. For our purposes, I especially find the overall Six Sigma process and how Six Sigma work is organized in companies to be of the highest relevance. For more detailed treatments of these and other aspects of Six Sigma, consult the following sources: Meisel, Babb, Marsh, and Schlicht‑ ing, 2007; Przekop, 2006; and De Feo and Barnard, 2004. 11.5.1  The Six Sigma Process The basic improvement process outlined in the Six Sigma concept is the DMAIC process: define, measure, analyze, improve, and control. You will certainly recog‑ nize facets of the plan, do, check, act cycle in this process, but a distinguishing fac‑ tor setting the DMAIC process apart is how systematic and rigorous it is. Another

Tools for Creating Improvements  205

key factor is that Six Sigma projects are selected by the company’s management in order to maximize payoff. This is completely in line with the overall methodology of this book, advocating the development of an overall improvement road map to ensure coherency among initiatives. The five phases of the DMAIC process focus on the following issues: • Define: The purpose is to scope the project by defining its focus, objec‑ tives, and desired benefits. Important outputs include a project charter, an understanding of the business process that causes the problems addressed, and insight into the stakeholders of the process. In this phase, many of the tools discussed earlier in the book are applied, including process mapping (where Six Sigma often terms a process map a SIPOC [suppliers, inputs, process steps, outputs, and customers] map), stakeholder analysis (some‑ times including Kano model requirements), and VOC (voice of the cus‑ tomer, a variant of QFD). • Measure: This can be compared to the phase for collecting data about the performance shortcoming presented in this book, with the purpose of nar‑ rowing down the range of potential causes of problems and understand‑ ing the capabilities of the current process. Achievements to be attained by the end of the measure phase include metrics for the process, that perfor‑ mance data have been collected and displayed to understand the variation of the process, and a calculated process capability. Typical tools used are check sheets for data collection; data presentation and analysis tools such as Pareto charts, histograms, run charts, and so on; and formulas for process capability. • Analyze: The purpose here is to better understand the problem: identifying possible causes, organizing these causes, and validating which are the root causes. Comparing again with the structure of this book, this is quite similar to the performance shortcoming analysis stage. Completing this phase of the Six Sigma project normally means producing a ­cause-­and-effect under‑ standing, creating a more detailed process map than the SIPOC chart, and verifying which causes have the most impact on the process. Tools for this stage include detailed process mapping, ­ value-­added analysis, the ­ cause­and-effect chart, the scatter chart, and more advanced hypothesis testing. • Improve: As the name clearly indicates, this phase of the Six Sigma proj‑ ect aims to implement improvements that remedy the identified problems, mirroring the phase of creating improvements in our overall improvement model. Outcomes from the improve phase are a ­ cost-­benefit analysis for the suggested improvement, a risk analysis for the implementation of the solution, an implementation plan, often a pilot implementation to verify the

206  Chapter Eleven

solution, and finally an implemented solution to the problems addressed by the project. Typical tools applied are different idea-generation techniques, ­cost-­benefit analysis, stakeholder analysis, force field analysis, and project management techniques. • Control: The purpose of this phase is to ensure that the process does not revert to its old state and to make sure that the solution actually works. Completing this final phase means documenting the results of the improve‑ ment, creating a plan for sustaining the improvement, devising a training plan to ensure minimal variation in the new process, and documenting the entire project. This normally relies on tools such as control charts, process capability calculations, and techniques for standardization. From this presentation, it should be quite obvious why many, including me, are reluctant to label Six Sigma a tool. I should also mention that the DMAIC process is not the only overall Six Sigma process model that exists. You might be familiar with the term design for X, where the X can be replaced with words such as manu‑ facturing, logistics, ­eco-­friendliness, and so on. Another replacement for X could be minimal variation, which would then translate into design for Six Sigma. For developing new products or for completely overhauling processes, there is in fact a technique called Design for Six Sigma (DFSS). Under this heading, an alternative to DMAIC has been developed: DMADV (define, measure, analyze, design, and review). DMAIC and DMADV are quite similar, so I will not present the details of this alternative process model. Common to both is that each phase is followed by a review of the outcomes. These are called toll gate reviews and are designed to ensure that one phase has produced the required platform for the next before the team is allowed to move on. Such reviews are held with people external to the project team. In Six Sigma, these people have special titles, which are described in the next section. 11.5.2  Six Sigma in the Organization To ensure that Six Sigma is carried out according to the rigor prescribed, an organi‑ zational structure has been designed to surround the individual Six Sigma projects. The main elements and roles in this structure are: • Careful project selection, as was already briefly mentioned, to ensure that the best projects are selected and that the projects target critical stakeholders and match the organization’s goals. • Black Belts and Master Black Belts, people who have undergone careful training in the Six Sigma methodology and are thereby qualified as project managers. They typically spend all their time running Six Sigma projects, but the position is not permanent. A selection of those certified as Black

Tools for Creating Improvements  207

Belts can be selected for the position of Master Black Belt, a highly quali‑ fied Six Sigma consultant. Master Black Belts teach Six Sigma inside the organization and work as consultants on projects. • Green Belts, employees who have received some training in Six Sigma and work part-time on projects (typically 20–25 percent of their time). They can even lead their own Green Belt projects, which are typically inside a single department. • Other team members, who are employees without Black or Green Belt cer‑ tification. There are rarely enough Black or Green Belts to fill all the posi‑ tions on the Six Sigma project teams, so other people are enrolled, either because they know the process well or because of personal skills. • Project sponsors, senior or middle managers who act as Champions or sponsors of individual projects. The sponsor, among other responsibili‑ ties, ensures that the team has the required resources, helps set the strategic direction of the project, and participates in the project reviews. • Six Sigma steering committee, the primary driver of Six Sigma deploy‑ ment in the organization. Larger companies can have more than one steering committee. The committee responsibilities include project and team selec‑ tion and preparing the organization for Six Sigma. There are other roles in the overall Six Sigma scheme, but these are the most important. If you are contemplating initiating a Six Sigma program in your organi‑ zation, I suggest you view this brief Six Sigma description as an introduction and consult more extensive sources before proceeding. In terms of an example of Six Sigma, you have already seen that the overall Six Sigma process is quite similar to the overall improvement model advocated by this book. A full Six Sigma example would more or less mean covering most of the tools presented thus far. Therefore, I have not included one here, but instead refer you to the two more comprehensive cases presented in Chapters 13 and 14. Let us therefore move on to the concept of business process reengineering.

11.6  Business Process Reengineering Business process reengineering (BPR) blossomed during the 1980s and parts of the 1990s. In all fairness, the approach has been heavily criticized—among other reasons, for recommending reorganization as a quick fix for complex process prob‑ lems and for advocating downsizing and a ­cost-­reduction focus over seeing growth opportunities. Long after its heyday and with most of the strong criticism behind it, BPR has a core message that still seems relevant: business processes can be tweaked and slightly adjusted, or they can be completely rebuilt. Although BPR,

208  Chapter Eleven

like many other approaches, does not represent a unified method, it contains prin‑ ciples that support both ways. As have other popular tools, BPR has been defined in a number of different ways. According to Hammer and Champy (1993, 32): Reengineering is fundamental rethinking and radical redesign of busi‑ ness processes to achieve dramatic improvements in critical contemporary measures of performance, such as cost, quality, service, and speed. This definition focuses on the objectives of BPR, that is, the breakthrough improvements. This element is also present in a definition by Peppard and Row‑ land (1995, 20): BPR is an improvement philosophy. It aims to achieve step improvements in performance by redesigning the processes through which an organi‑ zation operates, maximizing their ­ value-­added content and minimizing everything else. This approach can be applied at an individual process level or to the whole organization. If you compare these two definitions with some of the other tools available, many differences are apparent, but so are some similarities. What primarily separates BPR from the tools presented so far is the strong emphasis on the breakthroughs or radical improvements. BPR is not the tool to select if the objective is to achieve a 10 percent improvement. Rather, it is a strategic tool that can be useful when the organization wishes to achieve dramatic improve‑ ments in its performance level. At the same time, this seems very close to the streamlining presented in section 11.1, that is, increasing the proportion of ­value-­adding activities in the processes. The fact is, BPR is a collection of more or less related tools. Unlike streamlining, they have been put together to form a structured process that, when coherently applied, can generate the desired breakthrough improvement. It is not the intention of this book to dwell on definitions of BPR. It should, however, be mentioned that there are two principally different approaches when it comes to how BPR is carried out. These two approaches are linked to the role existing business processes play when applying BPR: • Should the existing processes form the basis for the new, redesigned processes? • Should the existing processes be changed in the improvement phase, or should new processes be designed to replace them? • Should the organization simply start the reengineering with a clean sheet?

Tools for Creating Improvements  209

So far in this book, much emphasis has been put on understanding our existing business processes. Ignoring these processes is very risky, as it means ignoring the knowledge and experience that have accumulated over a long period of time. In addition, there is a risk of repeating the same mistakes. It is also a fact that very few organizations have succeeded in implementing completely new processes. On the other hand, it avoids the danger of becoming buried in too many details of the existing processes and therefore allowing the limitations incorporated in them to impact the new process and thus limit improvement. As for most matters in this area, there are no correct answers. Between extremes, there is an endless number of degrees to which the existing processes influence the BPR effort. The important issue is to strike a suitable balance between how things are done today and how they should be done in a perfect world. From the two extremes, two types of BPR can be defined: • Systematic reengineering, where the current processes are understood, documented, and analyzed in order to systematically create new and better processes • Clean sheet reengineering, where the current processes are scrapped and new processes are created from scratch through fundamental rethinking Systematic reengineering can in many cases be synonymous with streamlining. Because starting with a completely clean sheet can rarely be done in practice, when referring to BPR throughout this book, something more moderate is intended. As opposed to streamlining, BPR is not only about incremental improvements in existing processes. BPR must seek to combine the best from existing processes with new ideas about how the ideal process should look. This is quite close to the technique of idealizing, whose objective is to create an image of the ideal process. Thus, a true application of BPR will be a structured process where these two ele‑ ments, idealizing and streamlining, are combined. 11.6.1  A Process for Conducting BPR There are many alternative processes that describe a BPR project. Kubeck (1995) inspired the process presented in this book. As with most other improvement pro‑ cesses, this process can be divided into parts: planning, solutions generation, and implementation. Since by definition the solutions generated in BPR can radically deviate from the existing processes, the middle stage has been divided into two phases: one for the actual generation of change proposals and one to evaluate how these can be implemented. This gives a process spanning four phases, as depicted in Figure 11.24. Before each of these phases is described in more detail, a brief summary of the process follows:

210  Chapter Eleven

Strategic plans

External performance requirements 1. Planning

BPR project 2. Reengineering

Recommendations 3. Transformation

Approved solutions 4. Implementation

Reengineered process

Figure 11.24  The process for conducting a BPR project.



1. Planning, where the focus of the BPR project is selected, a team to carry out the project is formed, and, if possible, objectives for the project are defined



2. Reengineering, where, more or less based on the existing process, a set of techniques is employed to reengineer the process to a level that will result in significant improvements



3. Transformation, where it is determined how the reengineered process can be implemented with regard to the existing process, the need for invest‑ ments, training, and so on



4. Implementation, where the solutions generated and approved during the two preceding phases are implemented and the process changed

11.6.2  Planning There are four central tasks in the planning phase: • Select the process to be improved through BPR and consider its scope • Assess the possibilities for achieving improvements and establish targets

Tools for Creating Improvements  211

• Establish a project team to perform the work in the project • Produce a project plan for the BPR project If this were a book dealing exclusively with BPR, the planning phase would be described over several pages. However, the contents of this phase are highly related to what has already been covered in this book, so most of the tasks have already been discussed and performed in the overall improvement process. The first step, selecting the process to be improved, is based on the results from the improvement planning and the improvement road map. The next step is an introductory assessment of whether BPR really is a suitable tool for improving this process. It is necessary to consider which improvement targets are realistic and desirable. This will aid in securing the necessary resources, both time and money, that are allotted for the project. The next task of the planning phase is to establish a project team. Most of the tools presented in this book are meant for use in teams or groups of some form. Some general guidelines for forming such groups were outlined in Chapter 6. The last step in this phase is to produce a project plan. The planning phase of the BPR project actually fits any preparation phase regardless of what tool is being used. This description is not unique to BPR, but it can be used in any improvement proj‑ ect. The ­BPR-­specific content appears in the next phase, the reengineering phase. 11.6.3  Reengineering The main purpose of the reengineering phase is, as the name implies, to reengineer the process to be improved. Before and after this main task, there is preparation and tidying up to do; thus, the steps of this phase are: • Document the existing process • Reengineer the process • Develop recommendations for improvement Whether choosing one or the other extreme (systematic reengineering or clean sheet), there is a need to know how the process is currently carried out. The first step of this phase is therefore to document the current process, usually by using a flowchart. If the book’s approach to improvement work has been fol‑ lowed, this job has already been done. This simplifies the work in this phase and makes it possible to start with the reengineering task. If much time has passed since the process was documented or if those participating in the BPR team are not familiar with this documentation, it might be necessary to update or repeat it.

212  Chapter Eleven

Turning to the reengineering of the business process itself, the procedure depends on the extent to which it is based on the existing process. Therefore, the two procedures are described separately. Systematic Reengineering As has been mentioned, systematic reengineering resembles the streamlining described in section 11.1. In BPR literature, the ESIA rules are often referred to. They are rules and techniques designed to accomplish four core steps in the sys‑ tematic reengineering of an existing process: • Elimination • Simplification • Integration • Automation Areas where each of these can be applied are listed in Table 11.3 (Peppard and Rowland, 1995). The ESIA rules should be applied in this order. The first step is to eliminate all activities that do not add value (see section 11.1). In organizations where process thinking is novel, a number of activities usually turn out to be ­non-­value-­adding and thus eligible for elimination without any negative consequences. Toyota’s rule Table 11.3  Focus areas for systematic reengineering. Eliminate

Simplify

Integrate

Automate

Excess production

Forms

Jobs

Dirty work

Waiting time

Procedures

Groups

Difficult

Transport

Communication

Customers

Dangerous

Processing

Technology

Suppliers

Boring

Storage

Problem areas

Data capturing

Defects/Errors

Flow

Data transfer

Duplication

Processes

Data analysis

Reformatting Inspection Reconsolidation

Tools for Creating Improvements  213

of thumb is that in many of the traditional manufacturing processes, 85 percent of the employees will be performing nonproductive work at any point in time (Griffiths, 1993, 20): • 5 percent can be observed not working • 25 percent are waiting for something • 30 percent are manufacturing something that increases inventory levels but does not add value • 25 percent are working, but in accordance with less efficient standards or procedures It is obvious that dramatic improvements can be achieved by attacking the areas listed under the “eliminate” heading in Table 11.3 and, of course, truly managing to eliminate them. When as many of these unnecessary tasks as possible have been eliminated, the next step is to simplify as many of those remaining as possible. Areas that are often overly complex are where the elements listed under “simplify” in Table 11.3 are generally involved. Thereafter, the remaining tasks are to integrate and further improve the flow of deliveries between the suppliers, the organization, and the customers. This can be accomplished at several levels. First, several smaller jobs can often be integrated into one single job. Such a job extension has several positive effects: it enriches the content of the individual employee’s work while at the same time eliminating interfacing between areas of responsibility. All such interfaces, no matter how well managed, create waiting time and potential quality problems. Second, different specialists can be integrated into groups that are responsible for more jobs. The closeness and autonomy created result in many of the same positive effects as the integration of several jobs into one. The third level is to attempt to integrate the organization with customers and suppliers. Organizations that achieve a particu‑ larly close relationship to their suppliers and customers usually experience more stable conditions and display better opportunities for common improvement. A typical application is to integrate customers and suppliers of core components into the product development processes. The final approach is automation. Information technology, computeriza‑ tion, and robotic equipment are support aids that can be quite powerful and give very good results if the processes being automated from the outset are sound. Automating problem processes can, in many cases, make things worse; that is, automation can enable errors to be made more rapidly. Therefore, it is essential to apply the principle of automation last, after elimination, simplifi‑ cation, and integration have been successfully applied. Processes suitable for automation are listed in Table 11.3.

214  Chapter Eleven

Automation usually fits the 80/20 rule. While many automation projects strive for 100 percent functionality in the automated system, a general rule is that 80 per‑ cent functionality can be achieved at 20 percent of the cost. Making systems that take into account every possible exception and special case can be very expensive. If a system that can handle most cases, with occasional manual intervention, is acceptable, both time and money can be saved. Clean Sheet It is difficult to give general advice about how to proceed with the clean sheet approach, that is, starting with a clean sheet and drawing an entirely new process. There are few, if any, specific techniques for this objective; we depend on creativ‑ ity, imagination, knowledge about available technology and human resources, and so on. Each BPR project utilizing this approach is unique. However, there are some basic questions that need answering: • Which basic needs do we want to fulfill, and for whom? By answering this question, we might realize that there are services or products that are pref‑ erable to those currently provided. A manufacturer of explosives asked this question, and the answer was the need for a hole in a rock, not the explo‑ sives per se. The customers would prefer to have the hole supplied without using the explosives. Therefore, the company started offering its services as explosive experts in addition to marketing the explosives, delivering fin‑ ished holes at the points indicated by the customers. • Why do we try to fulfill these needs? Does it fit in with the overall strategy of the organization? • Where must these needs be fulfilled? At home? In a separate service area? In the company’s facilities? The answer to this question can form the basis for dramatic changes in the processes. This happened in the restaurant industry, which discovered that not all customers wanted to eat at the restaurant, but would rather eat the restaurant meals at home. Thus, express delivery of hot food to private homes was created. • When must these needs be fulfilled? At what time of day and by which deadlines? Even if this is not a pure BPR project, many service institutions such as stores and banks have adapted their operating hours to meet the customers’ needs. This entailed many changes in their processes to accom‑ modate afternoon or evening needs. • How will the needs be fulfilled? What processes will be required? Who must perform these processes, and which technologies does this require? After answering these basic questions, we should start thinking about practical issues regarding how to satisfy these needs. That way, these basic questions are not

Tools for Creating Improvements  215

clouded by practical limitations. Just from looking at these questions and the chal‑ lenge the clean sheet approach poses, it is obvious that the creativity of the people involved is crucial. Following are some further questions to inspire the necessary creativity and imagination: • If you were to create a competitor of your own organization, how would it look to have the highest possible impact? • How should the ideal process look? • If we could build the entire organization from scratch, how would the orga‑ nization and this specific process look? These questions are summarized in Figure 11.25. Crazy ideas stemming from brainstorming not limited by practical considerations are combined with the answers to the preceding questions, ideas from benchmarking of other pro‑ cesses (described in section 11.7), available human resources, and technological opportunities. The last issue in the reengineering phase, no matter which approach is used, is to summarize the changes needed to implement the new process. By comparing on paper the newly designed process to the old process, we can decide which changes are required. This list of recommended changes will form the basis for the work in the transformation phase. 11.6.4  Transformation The purpose of the transformation phase is to build the foundation for an effective and successful implementation of the new process. The main tasks will be: • Evaluating the changes required to implement the new process • Planning the need for investments, training, purchases, and so on

Needs fulfillment: Customer requirements, demand patterns, when, where, practical limitations

Human resources

Process design

Benchmarking: Destroying sacred cows, what is actually possible, alternatives

Brainstorming, fantasizing, crazy ideas

Technological opportunities

Figure 11.25  Process design using the clean sheet approach.

216  Chapter Eleven

• Creating a favorable climate for change • Planning the implementation The reengineering phase created a number of recommendations for change without critically assessing their feasibility. The first step of the transformation phase is to apply a more sober perspective on things so as to evaluate whether all activities are possible to implement. The purpose is, of course, not to be so critical that most of the recommendations are rejected, but rather to ensure that infeasible changes or changes in opposition to the organization’s strategy are removed. For example, a recommendation that a company change its operating hours from 9 a.m.–5 p.m. to 2 p.m.–11 p.m. might be so radically against the human resource policy that it should be rejected. The next issue is to map the need for the changes and plan the necessary invest‑ ments, training, new staff, and so on. Many of the recommended changes will require these kinds of efforts, and the needs must be known before planning the implementation itself. If not, we risk, for instance, introducing the new process to operators who have not been sufficiently trained, as this requirement was never identified. Before the actual implementation starts, it is very important that the right cli‑ mate for change be created. Even the most ­ well-­planned improvement projects can be sabotaged if those who are impacted by the changes do not support them. Creating a favorable climate is a science of its own, one that is not covered by this book. It’s always favorable to have those who will later have to live with the changes take part in planning them. This means that including on the project team people already deeply involved in the process will create a good foundation for a painless implementation. Another issue is to always be open and honest about what is in the pipeline and provide a continuous flow of information to the affected parties. A department facing extensive change is like an area of low pressure in the atmosphere: it will suck up information of any sort until it is filled up. If this infor‑ mation is not constructive information from the management, it will be rumors and gossip that can be harmful to the situation. Force field analysis, as treated in section 12.3, is a suitable tool for creating the desired climate. The last step before the implementation phase is planning. The need for this plan is similar to the need for a plan for the BPR project itself, and it must contain approximately the same elements, in addition to the required investments, training, and so on. 11.6.5  Implementation Like the planning phase, the implementation phase of the BPR project, where changes are effected, is quite general. Regardless of which improvement project is discussed, the last activity is always implementation of improvements. Although

Tools for Creating Improvements  217

the changes resulting from a BPR project are often more comprehensive than when other tools are used, this phase in the BPR project is not in any way unique. At the same time, there is no doubt that implementation can be the most difficult phase. Many improvement projects that have performed excellently to this point fail in implementation. The implementation phase will not be described in detail here, but it will be described more generally in Chapter 12. The steps of this phase are as follows: • Set targets for the improvements • Create a favorable climate for change • Carry out the implementation plan • Monitor the progress of the implementation and handle any deviations Many talk about the two tools of BPR and benchmarking simultaneously, and they are to some extent related. This is treated in the next section.

Example As an example of the use of BPR, we will look at a manufacturer of hydrau-­ lic hoisting equipment for industrial applications. This company produced some standard systems, accounting for about 60 percent of its sales. At the same time, the company offered ­tailor-­made solutions for all potential purposes. Its reputation in the market was that it could do almost anything. Thus, for almost half of the sales, some degree of redesign of standard systems or development of new products and engineering of new solutions was required. The company was one of the few in the market offering ­tailor-­made sys-­ tems, but lately, several competitors had moved into this segment. Some competitors could offer shorter delivery times, even when the delivery involved product development, which endangered the competitiveness of our company. It was therefore decided to undertake a BPR project with the objective of reducing the delivery times for such systems by 50 percent.

Planning The process to be improved was identified, and a project team was established that consisted of two people from each of the following departments: • Sales • Manufacturing

218  Chapter Eleven

• Product development/projects (consisting of hydraulic and mechan-­ ical design) • Procurement The team decided to make a project plan and expected a ­six-­month dura-­ tion. The clean sheet approach was selected with consideration to the cur-­ rent situation. Thus, the intention was to combine the best from the existing process with elements of the ideal process.

Reengineering Following completion of these introductory tasks, the next step was to document the current process, as shown in Figure 11.26. A common denominator for the entire process was sequential develop-­ ment. First, the hydraulic department designed the core system functional-­ ity. Next, the mechanical department designed the mechanical system to fit the hydraulic system. Then the manufacturing department had to find a way to manufacture this system, while the procurement department was left to quarrel with the suppliers about components and processes for these. Often, the final product was different from what the customer had specified. On the positive side, the product development department used a very rational process to make drawings, system charts, ­bill-­of-materials, and so on. The core task of the BPR project, which was viewed with great expecta-­ tions, was to describe the ideal process. This task was carefully planned by: • Involving people known for their creativity and enthusiasm at work • Reserving two days per week for four weeks exclusively for this task • Creating in advance a climate of trust and cooperation within the group through two social gatherings • Ensuring the availability of all necessary aids for the job Without describing the group’s work in detail, the result, shown in Figure 11.27, was an outline of an ideal process that all agreed represented a dra-­ matic improvement over the current situation. In this process, parallelism was emphasized using ideas from concur-­ rent engineering and cooperation between internal departments and exter-­ nal elements such as customers and suppliers. The project group found this ideal process to be so powerful that it wanted to implement as much of it as possible. The reengineering phase was therefore concluded by sys-­ tematically summarizing all changes required to reach this target.

Transformation As the transformation phase was started by employing a more sober atti-­ tude, talks with the customers and suppliers revealed that while most of

Mechanical modification by the mechanical department

Manufacturing

Hydraulic design by the hydraulic department

Manufacturing

Hydraulic modification by the hydraulic department

Negotiations with suppliers and purchasing

Development of system functionality by the project department

Negotiations with suppliers, possible development of new components, and purchasing

Delivery and possibly installation

Mechanical design by the mechanical department

Delivery and possibly installation

Generation of process plans by the production preparation department

The project department assesses the need for redesign or product development

Figure 11.26  The current process for developing a tailor-­made system.

The sales department specifies the system

Customer request

Delivered system

Generation of process plans by the production preparation department

Delivered system

Production planning

Product development

Type of project?

Production planning

Modification project

Tools for Creating Improvements  219

Delivered system

Delivery and possible installation

Extension of the integrated team with a representative from the involved suppliers

Manufacturing

Figure 11.27  The ideal process.

Joint production planning between the manufacturing and procurement departments and the suppliers

Parallel development of hydraulic and mechanical system and process plans by the integrated team

Parallel development of hydraulic and mechanical system and process plans by the integrated team Purchasing

Joint production planning between the manufacturing and procurement departments

Customer request

Establishment of an integrated team of the customer and one person from the project, manufacturing, and procurement departments

Preparatory discussions between the customer and the sales department

Agreement on the supply of components between the procurement department and the suppliers

Product development

Type of project?

Modification project

220  Chapter Eleven

Tools for Creating Improvements  221

them were interested in cooperation, three central suppliers did not want this solution. Thus, the involvement of suppliers in the integrated team was scrapped. No other obstacles were found against organizing such proj-­ ects using integrated teams. No investments were required, but training was necessary in the areas of cooperation, group processes, and effective communication. Because central people from the affected departments had partici-­ pated in outlining the new process, the climate for change was favorable from the outset. To reinforce this climate, two ­one-­day information semi-­ nars were planned where the project group, in cooperation with external resource people within concurrent engineering, would present the project and its advantages. An implementation plan was also produced during this seminar.

Implementation Implementing the new process took nine months longer than expected. The results, however, were well worth the effort. The average time from when the first customer request was received to when a complete system had been developed, manufactured, and delivered was reduced from 4.5 months to less than two months. In addition, the climate in and between departments was significantly improved.

11.7  Benchmarking Benchmarking, which has been mentioned numerous times in the book, is a versa‑ tile method whose core element is comparison. It can be used for many different purposes: comparing your performance with that of competitors to identify per‑ formance gaps, comparing regional service offices’ customer satisfaction perfor‑ mance to determine bonuses, or comparing universities to present rankings of the best schools in different areas. Benchmarking and the key principle of comparison can be utilized at many different stages in the improvement work. However, in this chapter, I include it as an improvement tool. Before the actual process for conducting benchmarking is introduced, some definitions and other introductory issues related to benchmarking will be presented. The bulk of the description of benchmarking is based on Andersen and Pettersen (1996). 11.7.1  Definitions of Benchmarking The original meaning of the noun benchmarking is a predefined position, used as a reference point for taking measures against. This word has migrated into the business world, where it has come to mean a measured ­best-­in-class achievement recognized as the standard of excellence for that business process.

222  Chapter Eleven

As a verb, benchmarking can be somewhat philosophically defined as follows: Benchmarking is the practice of being humble enough to admit that some‑ one else is better at something, and being wise enough to learn how to match them and even surpass them at it. (APQC, 1993) This definition captures the essence of benchmarking, namely, learning from oth‑ ers. The term benchmark probably comes from geographic surveying, before the GPS era, where points in the terrain were given with reference to a fixed point, often a tall peak or some other easily recognizable and definable point. At that point, two stone slates were raised on end and spaced somewhat apart. A third slate was laid on top of these, thus forming a stone bench—hence the word benchmark. Because of its meaning of comparison against a reference point or the optimal, it was put to use in the business world, was extended, and can be defined as shown in Figure 11.28. Benchmarking, like BPR, emerged as a globally known tool during the 1980s. Like many other new tools, benchmarking was in fashion for a few years and was used and abused for all kinds of purposes. Fortunately, the tendency to view bench‑ marking as the answer to any issue has passed, and like BPR, benchmarking is now a tool just like many others, with its strengths and weaknesses. There has also been an evolution over the years in what we view as the strengths of benchmark‑ ing. Many used to see benchmarking as a method for comparing only key figures, often financial key figures, for the purpose of ranking the organization in relation to competitors or the industry average. This might have been the main application of benchmarking in the past, but today it is a far more powerful tool that is much more widely applicable. The core of the current interpretation of benchmarking is:

Comparing business processes, not just performance measures

A structured process

Learn from others

“Benchmarking is the process of continuously measuring and comparing one’s business processes against comparable processes in leading organizations to obtain information that will help the organization identify and implement improvements.”

External focus

Figure 11.28  Operational definition of benchmarking.

Change, not evaluation

Tools for Creating Improvements  223

• Measurement of your own and the benchmarking partners’ performance level, both for comparison and for registering improvements • Comparison of performance levels, processes, practices, and so on • Learning from the benchmarking partners to introduce improvements in your own organization • Improvement, which is the ultimate objective of any benchmarking study There are four main reasons for advocating the use of benchmarking in improve‑ ment work:

1. Benchmarking helps the organization understand and develop a critical attitude toward its business processes



2. Benchmarking encourages an open attitude toward seeking and sharing information and thereby is an active learning process that motivates change and improvement in the organization



3. Through benchmarking, the organization can find new sources of improve‑ ment and new ways of doing things outside its own environment



4. Through benchmarking, reference points are established for performance measurement of business processes

Thus, benchmarking is about comparing your organization with other organiza‑ tions. Different types of benchmarking can be defined depending on who you compare against and what you compare. First, you can compare on three different levels (at least) in terms of who you use as benchmarking partners: • Internal benchmarking—comparison against the best within the same orga‑ nization or corporation • Competitor benchmarking—comparison against the best direct competitors • Generic benchmarking—comparison against the best, regardless of indus‑ try or market The farther down the list we go in these types of benchmarking, the farther away from our own little world we move. This is illustrated in Figure 11.29. Because gaining access to competitors for benchmarking can be difficult, turning to other industries might be wise. At the same time, the chances are often higher that an organization in a different sector has developed approaches and business processes that are different from yours.

224  Chapter Eleven

Other industries Competitors Own company

Figure 11.29  Benchmarking out of the box.

On the other hand, identifying relevant benchmarking partners from other industries often requires significant creative leaps. The key is to think of someone with a similar challenge who is in a different setting. When successfully made, such novel links can cause breakthroughs for a whole industry. Some examples are shown in Table 11.4. Depending on what you compare, there are again at least three levels of bench‑ marking (the differences are illustrated in Figure 11.30): • Performance benchmarking—comparison of pure key figures or other perfor‑ mance measures. In sports, a high jumper doing performance benchmarking could learn where to set his targets from what the best in his age group achieved nationally, but he would learn nothing about how to make that height. • Process benchmarking—goes beyond performance measures by comparing how business processes are performed as well as how well they are per‑ formed. For the high jumper, this could answer questions like which jumping technique to apply, which equipment to use, which training methods to fol‑ low, what dietary issues to be aware of, and so on, to clear this height.

Table 11.4  Breakthroughs through benchmarking. Problem

Compared with

Long admittance times in hospitals

Hotel reception areas

Too lengthy setup of machines

Formula 1 pit crews

Planning the delivery of fresh concrete

Pizza delivery

Unstructured maintenance of power turbines

Maintenance of aircraft engines

Difficult to manufacture shell cases with the right cylindrical shape and smooth surface

Manufacturing of lipstick tubes

Tools for Creating Improvements  225

?

? ?

Figure 11.30  Three types of benchmarking based on what is being compared.

• Strategic benchmarking—comparison of strategic decisions and disposi‑ tions at a higher level. This is a less frequently used variant of benchmark‑ ing. In the sports example, this could perhaps teach the high jumper which competitions to enter to promote his career most effectively. Each type from these two categories—that is, who and what are being bench‑ marked—can, in theory, be combined into a benchmarking study with a given focus. In practice, not all combinations are equally suitable, as is shown in Figure 11.31. This is not explained in detail, but different studies have shown that the best results are generally achieved by a combination of process and generic benchmarking, using partners from other industries. The different types of benchmarking can be used in combination to some extent. A typical benchmarking study is outlined in Figure 11.32. As a start, performance bench‑ marking can be used to give an indication of where the organization stands and what Internal benchmarking

Competitor benchmarking

Generic benchmarking

Performance benchmarking

Process benchmarking

Strategic benchmarking

Relevance/value:

Low

Medium

High

Figure 11.31  Recommended combinations of types of benchmarking.

226  Chapter Eleven

High-level comparison: • Where are we? • How good are we in central areas? • Who is good? • Who can we learn something from?

Performance benchmarking

Benchmarking partners

Process benchmarking

List of prioritized improvement areas

Detailed comparison

of processes: • What do the best do? • What can we learn from them?

Figure 11.32  Progress of a typical benchmarking study.

should be improved, as well as to find possible benchmarking partners. Next, process benchmarking is used to improve through observing what the best do. The last definition issue relates to how benchmarking can be conducted with respect to the relationship to the benchmarking partners. The most common mod‑ els are shown in Figure 11.33. The most used model is where one organization initiates a benchmarking study with several other organizations as benchmarking partners. A ­one-­on-one comparison is made against each of these, often without letting the other partners know that more than one organization is being used and who the others are. It has become common practice for several organizations to join together to conduct a benchmarking study where comparisons are made among all the organizations involved. This usually gives better value for all partners involved. The third model, which is perhaps somewhat ­old-­fashioned, is where a third party,

One-on-one comparison

Comparison in a group

Anonymized comparison through a third party

Figure 11.33  Models for organizing a benchmarking study.

Tools for Creating Improvements  227

usually a consultant, performs the data collection and analysis and then presents the results to the organizations that contributed their data. As each organization cannot observe the best practices in use, it is often difficult to make any changes as a result of such studies. The model is best suited for performance benchmarking. 11.7.2  Conducting a Benchmarking Study Benchmarking is usually conducted in separate projects, normally called bench‑ marking studies, whose individual objective is to improve one of the organization’s business processes. In the larger setting of business process improvement, a bench‑ marking study typically forms one activity in the improvement stage. A full bench‑ marking study includes the activities necessary for the organization to: • Study and understand its own process • Find benchmarking partners • Study the benchmarking partners’ processes • Analyze the differences between its own and the benchmarking partners’ processes • Implement improvements based on what was learned from the benchmark‑ ing partners You will recognize some of these tasks as belonging to other phases of the improve‑ ment process. Typically, the first phase of studying your own process will have been conducted in earlier stages of the improvement project. Improvement implementa‑ tion also naturally falls within the next stage of the larger improvement process. There are a number of models that describe the different steps that constitute a benchmarking study; they differ in what to include and what to leave to other tools to handle. One such model is the benchmarking wheel, portrayed in Figure 11.34. Ordinarily, a benchmarking study lasts between six weeks and eight months, depending on the extent of the process being benchmarked, the number of bench‑ marking partners used, how data are collected from them, and so on. Often, imple‑ menting the improvements themselves, the main content of the adaptation phase, can take longer. For the other phases, approximately 50 percent of the time is spent on planning, 20 percent on observing the benchmarking partners, and 30 percent on analyzing collected data. A brief description of each of the phases follows. Plan The planning phase contains four basic activities:

1. Select the process to be benchmarked



2. Establish a benchmarking team

228  Chapter Eleven

5

1

Adapt: Choose “best Plan: Critical success practice,” adapt to factors: select a process for the company’s conditions, benchmarking, document the and implement changes process, and develop performance measures

4 2

Analyze: Identify gaps in performance and find the root causes for the performance gaps

Search: Find benchmarking partners

3 Observe: Understand and document the partners’ process, both performance and practice

Figure 11.34  The benchmarking process based on the benchmarking wheel.



3. Understand and document the process to be benchmarked



4. Establish performance measures for the process

These steps apply only if benchmarking is used as an isolated tool. Because this book encourages a more coherent process where improvement needs are identified at an early stage, we can proceed directly to the next step of the planning phase. Thus, no description is given on how to select a process for benchmarking. The ­benchmarking-­specific content appears in the search phase. Search From the benchmarking wheel, it might seem that the different phases should be performed sequentially, but this is not the case. If the entire planning phase is com‑ pleted before the search for benchmarking partners is started, the progress of the study will usually display a significant drop, as searching for partners usually takes some time. It is recommended that the task of searching for benchmarking partners be started in parallel with the planning, often quite early in the planning phase. The contents of the search phase are as follows:

Tools for Creating Improvements  229



1. Compile a list of criteria that an ideal benchmarking partner should satisfy



2. Search for potential benchmarking partners



3. Compare the candidates and select one or more partners



4. Establish contact with the selected partners and gain acceptance for their participation in the study

The first of these steps is actually a ­search-­technical task. No matter what is being searched for (for example, books in a library or partners for benchmarking), it is preferable to limit the search parameters. Starting a search among all possible organizations and considering them potential partners can result in an extremely complicated search. It is wise to consider the qualities the ideal benchmarking partner should possess, and consciously search for these. Typical issues included in such criteria are: • Geographic location • Size • Technology used and markets served • Industry • Structure and organization This list could be endless. However, it is important not to limit the search too much, either, as we might exclude partners that could be suitable, if not perfect, for the study. To find potential partners, a number of sources can be utilized. The organization’s own network can usually provide useful information pertaining to customers, suppliers, and others with whom there is cooperation. Different experts within the field can often be of help. Various industry associations possess knowl‑ edge about players in the field. Media coverage can give hints, as can all public information. By using carefully selected search keywords, vast amounts of infor‑ mation can be found on the Internet. In fact, only the imagination sets boundaries for the search, although this has its pros and cons. Since the information available is so vast, it is easy to become paralyzed, not knowing quite where to start. Many benchmarking study participants conclude that this phase is indeed difficult and should not be underestimated. It is also worth considering what performance level we want the bench‑ marking partners to be at. The pyramid in Figure 11.35 depicts the number of potential benchmarking partners available within each performance level. For world class, there may be at most only one candidate for each business process,

230  Chapter Eleven

while the number increases rapidly with decreasing ambition level. Thus, it can be much easier to find a moderately good benchmarking partner than holding out for someone of ­world-­class ability. A stepwise climb is more reasonable, as it might be close to impossible to go directly to the business processes found among the very best. Note that there might be many organizations using radi‑ cally different processes but not achieving any better performance levels. Do not mistake these different processes for better processes and fall into the trap of adopting them. After a number of potential benchmarking partners have been identified, compare and assess them before making a final selection. Often during this phase, a form of performance benchmarking will be carried out to determine how good the different contenders are. Generally, more than one partner is rec‑ ommended, the usual being between three and five. The costs increase with each additional partner, but they are outweighed by the advantages and the potential for finding best practices. The last activity in the search phase is to establish contact with the desired partners for the benchmarking study. There is no sure way to design such a query or to present it to ensure acceptance from the potential partners. An important ele‑ ment, however, is that you are willing to enter into a ­give-­and-take agreement. A willingness to open your organization and share the benchmarking information increases the likelihood of a positive response.

Best in class, world class

Best practice

Improvement compared to current performance

Equal to or lower than current performance

Figure 11.35  Ambition pyramid for benchmarking partners.

Different

Tools for Creating Improvements  231

Observe In the observation phase, the benchmarking partners’ processes are studied and documented, just as was done for our own process in the planning phase. The observation phase covers three steps:

1. Assess the information needs and information sources



2. Select a method and tool for collecting data and information



3. Perform data collection and debriefing

In benchmarking, we often say that information is sought at three levels, as shown in Figure 11.36. It is important to know the differences between these levels. The perfor‑ mance level indicates how well the organization performs a business process. To learn something and become better, it is necessary to go beyond the performance level and look at how the process is performed: the practices. These practices are specific process elements that can be implemented in our own situation to achieve improvements. There is no guarantee that a process will work in your organization. Therefore, enablers must be captured when studying the process of the benchmarking part‑ ners. Enablers are conditions surrounding the process that render the partner able to perform the process in a particular way. These can be IT systems used, contracts with suppliers, competence levels of employees, and so on. If these conditions are not captured, time and money will be spent attempting to implement a new process that cannot work due to lacking enablers. A number of methods are available for collecting information and data about benchmarking partners. We separate between methods (the means utilized to get in touch with the partner) and tools (the specific techniques used to collect data). The matrix of Figure 11.37 summarizes some of the most common tools and methods, and the Xs indicate relevant combinations. Each of these methods and tools is actu‑ ally a field of its own. The last step of the observation phase is debriefing, which happens after data have been collected from a benchmarking partner. After being away from the office for a

Enablers

Practice

Enablers

Figure 11.36  Three levels of benchmarking information.

Performance level

232  Chapter Eleven

Tools Methods

Questionnaire Interview

Postal investigation

X

Telephone

X

X

Partner visit

X

X

Direct observation

X

Figure 11.37  Methods and tools for the observation phase.

while because of this task, you will often have numerous urgent matters to deal with upon your return. If these matters are attended to before you process the collected infor‑ mation, many impressions and ideas may be lost. It is therefore important to set aside time immediately after data collection for a debriefing within the team in order to: • Collect and sort information and data • Transcribe interview notes • Take down mental impressions and other ideas not yet captured on paper Analyze After the information and data about the benchmarking partners’ processes have been collected, the next step is to analyze the data and identify improvement sug‑ gestions. The analysis phase consists of five steps:

1. Sort the collected information and data



2. Quality control the collected information and data



3. Normalize the data



4. Identify gaps in performance levels



5. Identify causes of the gaps

The first two are really preparatory tasks that are included to ensure that all the data are available and accurate. Before starting the actual analysis of the data, it can be helpful to normalize the data set. A common argument against benchmarking is that what is being compared is unique, so no comparison is even close to being relevant. Most things can, however, be compared if a normalization is carried out beforehand. Normalizing means adjusting for the conditions that are truly different. It often consists of recalculating the data to average values or ratios that eliminate aspects like size, market conditions, different legislation, cost levels, and so on. One central element of the analysis phase is a gap analysis. This entails iden‑ tifying gaps as well as determining the causes of these gaps. The gap itself is of little help when it comes to identifying improvement actions. The purpose of the

Tools for Creating Improvements  233

gap is to prove that one partner does something that makes him or her able to per‑ form better than another. The presence of a gap is thus more of a signal that there is something worth examining more closely. There are several techniques available for finding the causes of the gap. A rela‑ tively simple approach is to compare the flowcharts for the processes directly. This can provide much information about the differences that contribute to higher performance. Other tools that have already been described but that are suitable for this purpose are ­cause-­and-effect charts, relations diagrams, and five whys analysis. Irrespective of the approach employed, the goal of this phase is a list of the conditions believed to con‑ tribute to the superior performance level of the benchmarking partners. Adapt As with the planning phase, the adaptation or implementation phase is not really benchmarking specific. Implementation in general is treated in Chapter 12; thus, the description here will be very brief. The adaptation phase consists of four steps:

1. Describe the ideal process and summarize improvement actions based on it



2. Set targets for the improvements



3. Develop an implementation plan, carry out the plan, and monitor the progress



4. Write a final report from the benchmarking study

After the benchmarking study has been completed, a final report should be produced that covers the entire study and contains lessons learned and recommendations for future studies. This report will constitute documentation of the study to all parties involved and interested. Furthermore, it should describe improvement actions that this project did not implement. The document should be sent to the benchmarking partners in return for their participation. Finally, it should contain some record of all bench‑ marking partners involved and that could become involved in ensuing studies.

Example As an example of a successful application of benchmarking, a description of a study performed by the telecommunications company Pacific Bell is included. Pacific Bell is one of several companies offering phone services to the residents of California. It was separated from the AT&T monopoly when the market was deregulated. One means of survival in an increas-­ ingly competitive environment has been continuous measurement of cus-­ tomer satisfaction. After using this method since the early 1980s, Pacific Bell started a benchmarking study to look at and improve the system. The

234  Chapter Eleven

basis for the study was a general concern about whether the measure-­ ments were accurate and whether they were used sensibly. The bench-­ marking study is described here, phase by phase.

Planning of the Study First, a benchmarking team was established, consisting of: • The manager for the measurement department • Four employees from the measurement department, responsible for the areas of collection, analysis, and application of customer satis-­ faction data • The central benchmarking manager of the company, who acted as a facilitator of the benchmarking methodology The first task of the team was to document the current process for measuring customer satisfaction. A very thorough job was done in this respect, result-­ ing in a 17-page description of all processes related to this area. Besides this task, the team members participated in different conferences, seminars, and other events that focused on the topic of customer satisfaction measurement. The result was that the team was knowledgeable about its own processes as well as the field in general by the conclusion of the planning phase.

Search for Benchmarking Partners To find the best possible benchmarking partners, Pacific Bell defined a list of criteria for the organization and its customer service. For each element on this list, a value was assigned indicating its importance for a potential partner. In order of decreasing importance, the criteria related to the company itself were: • Profitability the last five years, as this was seen as an expression of the degree of customer satisfaction • Multiple market segments, to find partners operating in a similar situation to Pacific Bell’s • Service industry, as customer satisfaction measurement in the ser-­ vice industries was considered quite different from what was being done in manufacturing organizations • Long-term customer relationships, as opposed to companies depen-­ dent on ­one-­time sales, as ­long-­term relationships would enable a continuous measurement of customer satisfaction • Based in California, as this geographical area has a unique blend of people • Technology driven, to resemble Pacific Bell as much as possible • Changed regulation conditions, to see how such conditions impacted customer satisfaction and the measurement of it

Tools for Creating Improvements  235

The criteria defined regarding the customer service of potential partners were: • A leader in customer satisfaction • Active use of feedback from customers for process improvement • A quantitative and systematic system for measuring customer satisfaction • Using several different instruments for customer satisfaction data collection To identify potential benchmarking partners, one ­full-­time person was hired to conduct a very thorough search of literature to find companies match-­ ing the criteria. This resulted in a list of about 20 companies that seemed to match the profile. These 20 potential partners were divided among the team members to conduct an even more detailed assessment of each. From these evaluations, eight organizations were agreed on as objects for further study.

Information Collection The selected partners belonged to industries such as banking, telecommu-­ nications, insurance, and public relations. To collect information, a ques-­ tionnaire was produced that consisted of two parts, one quantitative and one qualitative. The quantitative part focused on figures for the number of employees who performed different tasks, costs, the number of customers followed up, response rates for measurements, and so on. The qualitative part was far more extensive and focused on how customer satisfaction measurement was performed, how it was applied, who measured it, what specific performance measures were used, and so on. Due to limited means for the benchmarking study, the team found it impossible to undertake the data collection through partner visits. Training was therefore given in interview techniques, and the questionnaire was answered by phone and fax. First, the companies agreeing to participate in the study received a document containing Pacific Bell’s own answers to all the questions in the questionnaire. Next, the quantitative part was filled in by the partner and faxed back to Pacific Bell. After examining the answers, the team handled the qualitative part over the phone.

Data Analysis To conduct the analysis in a sensible way—that is, by spending sufficient time on it—a number of ­full-­day meetings were agreed to. First, the quantitative data were compiled in a matrix and analyzed, which was a rather easy task. The analysis of the qualitative answers was far more complex, because all answers had to be discussed in plenary to create a full understanding of the partners’ processes. This phase led to a list of recommendations for Pacific Bell: • Establish an organization to handle all customer responses and use it to improve the products and services

236  Chapter Eleven

• Terminate the generation of customer satisfaction data on the level below managers, to avoid the employees’ fear of repercussions • Stop basing the payment for ­lower-­level managers on customer sat-­ isfaction data • Develop internal process indicators linked to customer requirements • Expand the scope of customer satisfaction measurement, but reduce the measurement frequency to once every three months • Survey both customers who had recently been surveyed and those who had not, while also trying to reach the customers who rarely give any feedback • Use customer satisfaction data at a strategic level • Eliminate the frustration of employees at lower levels in the organi-­ zation who are held responsible for measures they can only partially control

Adaptation and Improvement Within a month after having presented the recommendations, management agreed to reduce the extent of the questionnaire used for measurement. This led to savings of about $1 million, and thus immediately saved far more than the total costs for the benchmarking study. More importantly, actions were taken to understand in detail the custom-­ ers’ true requirements for the services delivered by Pacific Bell, as exemplified by the benchmarking partners. Furthermore, pilot implementations were under-­ taken of a totally new system for customer satisfaction measurement, where the responses were used only at a high organizational level, and the other employees had process level measures developed. A number of less exten-­ sive changes were introduced regarding measurement of customer satisfac-­ tion, related to specific processes for customer service. For instance, a system was established where immediately after the installation or repair is completed, the customer is notified—a practice that was not common before.

Achieved Results The costs for the study reached $70,000. The reduction in measurement extent and the introduction of a new measurement system for customer satisfaction are expected to give savings in the area of $5 million. Even more important than the savings is the fact that a number of processes were dramatically improved as a result of the study, which is expected to result in large gains in terms of increased customer satisfaction. In addition, many of the changes improved the work situations of many employees.

Having concluded this phase of the improvement project, you should be ready for the final leg: implementing the improvements.

12 Tools for Implementing Improvements

A

ll the previous categories of tools have to some extent been aimed at creat‑ ing solutions or improvement actions to solve different problems. If these improvements are not implemented in the organization, the entire effort is wasted. Excellent improvement suggestions do no good lying in a drawer or on a shelf; they must be brought to life. This very difficult task consists of several subtasks: • Sorting and prioritizing among the improvement proposals • Organizing the implementation • Setting targets for the improvements • Developing an implementation plan • Creating acceptance for the required changes and a favorable climate for the implementation • Carrying out the implementation itself This chapter partly addresses some general questions related to these tasks and partly describes some specific tools. The tools presented are: • A∆T analysis • Tree diagram and process decision program chart • Force field analysis The remainder of this chapter contains a blend of descriptions of the more general tasks in the implementation phase and the more common tools. The first decision to be made in the implementation phase of an improvement project is usually what to

237

238  Chapter Twelve

implement. Often, the work of the preceding phases will have generated a number of possible changes. Rarely will the resources required to implement all suggestions be available. It is therefore a natural first step to prioritize the improvement propos‑ als and start with those expected to produce the greatest effects or that for some other reason are preferred. Criteria that can be used for sorting the proposals are: • The investment needs for introducing a new method or process • The training needs for a new process • Time limitations, both in the form of deadlines for the project and organi‑ zational restrictions with regard to the time available for performing the implementation task • The organization’s motivation level (that is, burned out or still enthusiastic) A little more systematically, a ­four-­field matrix can be designed, using the follow‑ ing issues as determining factors: • Expected improvement effects • The “implementability” of the improvements, that is, how easy or difficult they will be to introduce in the organization This will give a matrix like the one shown in Figure 12.1. Improvements that fall within the ­low-­hanging-fruit area are obviously the preferred ones to start with. Such a sorting will give a prioritized list of improvement actions in the order they should be implemented. There are several options for deciding how the imple‑ mentation should be organized: • By the original improvement team, where the same team that has carried out the project so far also undertakes implementing the improvements. The advantage of this approach is that the team knows the project and what the solutions entail.

Expected improvement effect

Expected ease of implementation High

Low

High

Low-hanging fruit

High effect, difficult to implement; second choice

Low

Low effect, easy to implement; second choice

Wasted effort

Figure 12.1  Matrix for assessing implementation order.

Tools for Implementing Improvements  239

• By a specific implementation team, where a new team is formed that con‑ sists of the necessary and suitable people to assume responsibility for the implementation itself. Even if this team does not know the work so far, it is often wise to include people in the implementation who were not involved in the development of improvement proposals. Thus, this approach can be sensible in some cases. • In the line organization, where the functionally responsible people assume responsibility for implementing changes by using the resources of the ordinary organization. This is perhaps the most common model, where we ensure that those who will later perform the process also take part in implementing it. After the model for organizing the implementation is selected, the next logical step is to set targets for the improvements: which performance level we want to achieve after the implementation has been completed. For this purpose, A∆T analysis is a helpful tool.

12.1  A∆T Analysis A∆T is a tool that is quite closely related to both idealizing and ­value-­added analy‑ sis. Whereas both of these tools are used to create improvement proposals, the main purpose of A∆T analysis is to set ambitious targets for the improvement work. The method is based on the assumption that it is always possible to find two durations, accumulated costs, total numbers of defects, or other accumulated performance measures for a given process: • “A” stands for actual: the actual time, costs, and so forth, currently related to performing the process to be improved • “T” stands for theoretical: the theoretically fastest time, lowest cost, and so forth, that can be achieved when performing the process With A∆T defined this way, we see that the theoretical value is closely related to the ideal process in idealizing. Furthermore, if we are considering only time or cost, the theoretical value can often be found simply by subtracting OVA and NVA, identified through a ­value-­added analysis. The analysis can be used in conjunction with these two tools. When performing an A∆T analysis, these two values can be used in two different ways. First of all, we can calculate the ratio between the A and T values: A . T This ratio, ∆, expresses the improvement potential in eliminating all unnecessary activities and performing the process as efficiently as possible. The higher the ratio, the higher the potential. This ratio can also be used as an expression of how much there is to gain by approaching the ideal process.

∆=

240  Chapter Twelve

For setting improvement targets, there is no need to calculate ∆. Instead, the target can be based on the T value. Whether the target should be set equal to the T value or somewhat less ambitious to take into account any practical limitations is an assessment that must be done in each specific situation. The steps of the analysis are as follows:

1. Start the analysis from the flowchart for the current process.



2. In the flowchart, add figures for time, cost, number of defects, and so on, for each activity.



3. Critically evaluate each activity to determine whether it adds value. If it does not, determine whether it can be eliminated. Activities that can be eliminated are marked with a color or by some other suitable means.



4. Summarize the A values and the T values, where the T values are all the nonmarked activities, and calculate the ratio A/T.



5. Set the improvement target at or close to the T value.

Example A company that had reached the implementation phase of an improvement project had some problems in determining an improvement target for a reduced process duration. A suitable tool for this task was the A∆T analysis, as the company had produced a flowchart for the process and partly performed a ­ value-­added analysis. The flowchart with attached activity duration times and activities eligible for elimination marked is shown in Figure 12.2. As can be seen, the ratio between A and T is 2.05, that is, a potential for approximately halving the duration time of the process. To be a little more realistic, the target was reduced somewhat compared with the ideal, down to 12 days. 1.5 C 3.5 A

2.0 B

1.0 E 1.5 D

= can be eliminated

A 21.5 = = 2.05 T 10.5

3.5 F G 4.5

2.0 I 0.5 K

Figure 12.2  Flowchart for the A∆T analysis.

H 2.0 3.0 J

Tools for Implementing Improvements  241

When it comes to improvement targets, or targets in general, they should be: • Ambitious enough to require some effort to be reached. Targets that are too easily attainable do not pose any challenges and rarely inspire great motiva‑ tion. The result can therefore be a lower improvement gradient than what could have been the case had the targets been more ambitious. • Realistic, so as not to deter. Targets that are too ambitious can cause the pen‑ dulum to swing too far to the other side, leading to frustration and reduced effort. • Operative, so they are easy to comprehend and follow up, to monitor whether we are approaching the target.

Example It has been claimed that when Ronald Reagan won his first presidential election, he gave all federal offices instructions to increase productivity by 5 percent within one year. After a year had passed, the results started com-­ ing in to the administration office, and most of them were between 4.8 and 5.2 percent productivity increases. The target of 5 percent was probably not sufficiently ambitious, and a target of 10 percent would probably have resulted in figures between 9.8 and 10.2 percent. Because the target was set too low, it did not inspire and motivate sufficiently to reap the full effects possible. If, on the other hand, the target had been set at 50 percent, the results still might have been in the vicinity of 5 percent, as this would have been asking too much, creating frustration.

12.2  Tree Diagram and Process Decision Program Chart Any project should have a project plan, as has already been emphasized. This is also true for improvement projects, especially in the implementation phase of such a project, which can often be viewed as a project of its own. Implementing improvements can often take longer than the total duration of the project thus far. It is therefore essential to have a project plan to guide this activity. Generally, such a plan should cover the following elements: • Activities—those tasks that need to be carried out to implement the improve‑ ment proposals • Sequence—the order in which the activities must be carried out • Organization and responsibility—an indication of who is responsible for each activity, both carrying it out and monitoring its progress

242  Chapter Twelve

• Schedule—a more detailed plan for when the activities should be carried out, including milestones for central results expected throughout the project • Costs—estimates for the costs involved in the implementation There are several techniques for project planning that involve different levels of scope and complexity. One ­easy-­to-use tool, suitable for breaking down larger tasks into activities of manageable size, is a tree diagram. This diagram can be combined with more complicated calculation methods for the project, for example, PERT (Program Evaluation and Review Technique) or CPM (Critical Path Method). The approach for creating a tree diagram is as follows:

1. Generate a list of activities that must be performed to implement the improvement proposals



2. On sticky notes, write down each activity in the form of a verb followed by a noun



3. Arrange the activities in logical subgroups that must be performed in sequence



4. Arrange the subgroups in an overall sequence to illustrate the entire plan in the tree diagram

A typical diagram will look like the one shown in Figure 12.3. The result will typically be a hierarchy of activities. The main activities, read from left to right in the diagram, represent the main tasks of the implementation. These main tasks will often be artificial activities (that is, names of a group of subactivities). Each of these main activities will therefore include a number of subactivities below them, to be carried out in the order they are presented from left to right. For each activity in the diagram, information about deadlines, responsibility, costs, and so on, can be attached.

Implementation objective

Main activity 3/15 1

Main activity 6/15 2

Main activity 10/1 3

Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity Subactivity 1.1 1.2 1.3 2.1 2.2 2.3 3.1 3.2 1/1

2/15

3/15

Figure 12.3  A principal tree diagram.

Tools for Implementing Improvements  243

Example A library decided to introduce a new ­computer-­based registration system. To plan this task, the employees designed the tree diagram in Figure 12.4. For each activity, the completion date was attached.

Introduce computer-based registration system

Buy the system

Invite bids

Evaluate bids

1/1

2/15

Train the staff

3/15

Negotiate Buy training Run contract package courses 3/15

4/1

6/1

6/15

Put the system 10/1 into use

Test the skills

Transfer old data

Start the system

6/15

9/1

10/1

Figure 12.4  Tree diagram for introducing a computer system.

The plan described in the preceding example is not very detailed and does not consider any unforeseen events during the implementation. This is typical in project plans of this type: they do not say what to do if something goes wrong. To include this element in the planning, we can use a process decision program chart (PDPC), which is really an extension of the tree diagram. This is a planning tool for making detailed implementation plans that include all possible negative events and prob‑ lems that could occur along the way. Predicting such problems before they occur makes it possible to address them. This enables some form of preemptive problem solving, which is inexpensive compared with starting to think about solutions only after the problem has occurred. The PDPC is most often used when a large and com‑ plex task is to be carried out for the first time, where the costs associated with failure are exceedingly high, and where finishing by the deadline is critical. The approach for using PDPC is as follows:

1. Generate a tree diagram for the implementation task or use one that has already been designed. At least to start with, it is wise to use a tree diagram that is not too complicated, as this can require undue time for analyzing possible problems for small, unimportant activities. An appropriate com‑ plexity contains the main activities with one level below them, as in the example shown in Figure 12.3.



2. For each element at the lowest level of the tree diagram, ask questions like, what potential problems could occur during this activity? and what

244  Chapter Twelve

could go wrong here? For these questions, brainstorm a list of answers for each potential problem area. When no more answers surface, examine the list to eliminate problems that are unlikely or that are expected to have no significant consequences. Each element should include an assessment of the consequences in terms of time, cost, and quality.

3. Add the remaining potential problems, those that are considered sig‑ nificant, to the diagram as ­what-­if elements below the lowest level of activities. Use a different color or shape to separate these elements from the activities.



4. For each ­what-­if element, brainstorm possible countermeasures that can be undertaken if the problem occurs. These countermeasures should consist of reserve activities and indications of duration and cost. Place all countermeasures in the diagram, which is being transformed from a tree diagram to a PDPC. Place the countermeasures under the ­what­if elements and link them to the potential problems they solve. Again, use a different color to separate them from the ­what-­if elements and the activities.



5. Finally, evaluate each countermeasure with regard to ease of implementa‑ tion, practicality, effectiveness, and so on. Mark difficult or ineffective ones with an X, and mark those you expect to be effective with an O.

A general PDPC is shown in Figure 12.5. Designing such a chart forces us to anticipate all possible problems that could occur during implementation and, Implementation objective

Main activity 1

Main activity 2

Main activity 3

Subactivity 1.1

Subactivity 2.1

Subactivity 3.1

What if 1.1.1

What if 1.1.2

What if 2.1.1

What if 2.1.2

What if 2.1.3

What if 3.1.1

Countermeasure A

Countermeasure B

Countermeasure C

Countermeasure D

Countermeasure E

Countermeasure F

X

O

O

O

X

O

Figure 12.5  A general PDPC.

Tools for Implementing Improvements  245

importantly, develop countermeasures. This might change the original plan in order to avoid potential problems as well as prepare countermeasures in case problems do occur.

Example The library that designed a tree diagram for introducing a new ­computer-­based registration system took its study even further and created a PDPC for the imple-­ mentation plan. The chart is shown in Figure 12.6.

Introduce computer-based registration system

Buy the system

Invite bids

Evaluate bids

No bidders

All bids too expensive

Cover entire U.S. O

Put the system into use

Train the staff

Negotiate contract

Buy training package

Run courses

No packages available

No free time

Test the skills

Transfer old data

Hard to learn Wrong how to use data format

Make own system

Have tailormade

Use overtime

More training

X

O

O

X

Adapt the data O

Start the system

System crash

Trial operation O

Figure 12.6  PDPC for the library.

12.3  Force Field Analysis As mentioned at the beginning of this chapter, an important subtask in the implementa‑ tion phase is to create acceptance for the suggested changes and a favorable climate for their implementation. This is a significant task that involves disciplines such as psychol‑ ogy and human resources management. A change process can be viewed as a formula involving the following elements: • Q = quality of the change approach • A = acceptance of the change among those involved • E = effectiveness of the change process

246  Chapter Twelve

The formula is:

E = Q × A.

Studies have shown that all successful change processes have high values for both Q and A, while most failed change projects also have high Q. The shortcomings are related to the acceptance of the changes. According to Miller (2002), there are often six layers of resistance: • Disagreement that there is a problem at all • Thinking the problem is outside our control • Argumentation that the suggested solution cannot solve the problem • Argumentation that the suggested solution will also cause negative effects • Creation of barriers against the implementation • Creation of doubt about whether others will cooperate for the solution A general piece of advice is that the more information given to those who will be affected by the changes, the less resistance will be met. Thus, communication with everyone affected and others who might represent obstacles to an effective imple‑ mentation is essential. This includes: • Top management, which has the authority to decide on implementation and allot the necessary resources for it • Everyone involved in the process to be changed, as it is essential to motivate them for the change • Everyone delivering input into the process or receiving output from it, as they could also be affected by the changes • Other gatekeepers or people who can impact the implementation and its progress, usually people with financial authority When it comes to creating a positive climate for the ensuing changes, force field analysis is a helpful tool that can contribute to creating an overview of the situa‑ tion and possible actions to improve it (Andersen and Pettersen, 1996). Force field analysis is based on the assumption that any situation is a result of forces for and

Tools for Implementing Improvements  247

against the current state that are in equilibrium. An increase or decrease in the strength of some of the forces will induce change—a fact that can be used to cre‑ ate positive changes. A similar method is often referred to as benefits and barriers analysis. The approach is almost identical to force field analysis, with the excep‑ tions that forces for the change are called benefits and forces against are called barriers. This method does not involve a diagram. The procedure for using force field analysis is as follows:

1. Clearly define the change desired. This is information that can usually be taken directly from the implementation plan and its improvement objectives.



2. Brainstorm all possible forces in the organization that could be expected to work for or against the change.



3. Assess the strength of each of the forces and place them in a force field diagram. The length of each arrow in the diagram expresses the strength of the force it represents.



4. Consider actions that could increase the forces for the change and reduce those against it, especially the stronger forces.

With these actions, the power balance surrounding the change could smooth the implementation. A force field diagram is shown in Figure 12.7.

Objective

Forces for

Figure 12.7  Force field diagram.

Forces against

248  Chapter Twelve

Example After identifying potential problems facing the introduction of a new computer system for registering transactions, as well as countermea-­ sures for these, the library undertook a force field analysis to understand the climate for this major change. Important forces for the implementa-­ tion included: • The new system would dramatically improve the process of check-­ ing books in and out of the library, making it easier on employees • The system would automatically generate reminders for overdue books, thus replacing a currently manual job • Searches for titles would be quicker and easier • If the system worked as anticipated, it would dramatically improve customer service, which currently had a somewhat tarnished reputation Several forces working against the change were also identified: • A general skepticism toward computer tools among many of the employees, whose average age was 42 • Fear of computer problems that could not be solved as easily as with manual systems • Fear of becoming redundant due to reduced manpower needs Put into a diagram with the forces’ strength represented by the arrow length, the resulting picture of the situation is as shown in Figure 12.8.

Introducing a computer-based registration system

Forces for

Simplified transactions

Automatic reminders

Quicker search

Forces against

Computer skeptics

Computer problems

Redundancy fear

Better service

Figure 12.8  Force field analysis for a new library computer system.

Tools for Implementing Improvements  249

For each of the forces, several measures were identified that could point the library in the right direction: visits to other libraries where com-­ puter systems had already been introduced with good results, general information and training in computer use, and so on.

Finally, some general advice about effective implementation: • To ensure full support for the changes, involve everyone responsible for results from the process being improved • Try to elicit inspiration from those involved in the project • Follow a clearly communicated plan • Keep the affected employees informed about progress and achieved results • Emphasize the importance of patience—changes do not happen overnight Perhaps the greatest challenge in this phase of the improvement project is to main‑ tain the team’s intensity to ensure that the implementation is carried out. Therefore, continuous monitoring of progress is necessary with regard to: • Time spent compared with the planned duration • Resources consumed compared with the implementation budget • The quality of the results

13 Sample University—Improving Student Satisfaction

U

nfortunately, it is not easy to find good working examples of the toolbox in a coherent improvement process. As with the literature in this field, there are a number of good case descriptions that portray the use of one or more isolated tools, but very few that illustrate the interplay between the tools. Therefore, this chapter contains a description of an organization, Sample Univer‑ sity, and its systematic use of a coherent system for process improvement. The description is based on a real case, but parts of the story have been embellished to demonstrate the use of different tools in a larger context.

13.1  Description of the Organization Sample University is a small university, with about 3000 students at the bachelor, master, and PhD levels. Courses of study include social sciences, engineering sci‑ ences, and management subjects. Sample University is divided into six faculties, each containing three to six departments. The university has about 150 professor‑ ships, close to 400 other faculty positions, and a technical/administrative staff of about 600 employees. The number of applicants to Sample University has been quite stable, although it has displayed a small increase in the past year. There is, however, an increasing pro‑ portion of students who choose shorter studies or who quit or transfer before complet‑ ing their degrees. Sample University receives some funding from the state as well as some support from industry and private donations, but the bulk of its income comes from tuition fees. The university is not among the ­highest-­ranked schools in the United States within its areas of study, but it holds a good position above the average. The immediate challenge was to drastically increase student satisfaction and to improve the university’s ranking in order to increase the ­long-­term number of applicants. This chapter describes how Sample University handled this challenge.

251

252  Chapter Thirteen

13.2  Development of Performance Priorities For years, Sample University had based its operations on a performance measure‑ ment that included: • National ranking performed by different agencies that was based on many aspects of the universities • Development in the body of applicants • Study completion percentage • Separate course evaluations From these performance measures, it was difficult to prioritize the areas or pro‑ cesses of desired improvement. At the same time, there was a wish to start with a pilot project because of the limited ­short-­term resources for this work and because this type of project would help individuals become familiar with the improvement process. To help prioritize the resources for the pilot project, a better set of perfor‑ mance measures was needed. Ideally, a comprehensive documentation of important processes would have been undertaken, but the pilot project and the preparations for it did not allow for this. The university’s management team, consisting of core academic and administrative personnel, decided therefore to conduct a measure‑ ment of the customer satisfaction of a selection of students. The management team wanted to combine this with input from companies that had hired graduates of Sample University over the last two years. A questionnaire was designed and distributed to about 200 students from dif‑ ferent departments. Concurrently, telephone interviews with both alumni and man‑ agers from approximately 20 companies in the area were conducted. There were eight main areas of evaluation:

1. Usefulness of the academic content of the studies



2. The faculty’s ability to convey the academic message



3. Routines for admissions and other administrative student policies



4. Housing and leisure facilities on campus



5. Pedagogical tools used in teaching



6. Tuition fees and other study costs



7. Relevance of the education for later jobs



8. The students’ abilities to cooperate in ensuing work situations

Sample University—Improving Student Satisfaction  253

Even though some variation was seen at the start among the faculties, average evaluation scores were calculated for the whole of Sample University. Because the competition among universities is fierce, the management wanted to form an impression of the situation compared with other colleges. Many universities col‑ lected and published corresponding data in study catalogs and similar material. From such sources, Sample University collected data from other universities in the state as well as from national competitors. The performance levels were compared using a spider chart, among others, as shown in Figure 13.1. In some areas, the results were far from encouraging. The areas displaying the largest gaps were: • Ability to convey the academic message • Pedagogical tools • Administrative routines • Students’ ability to cooperate in later work positions The results did not indicate what areas should be focused on, so Sample University used a performance matrix to prioritize the areas with a gap. The performance matrix is shown in Figure 13.2.

State B

Academic content

Ability to cooperate

National A National B Ability to convey message State A

State C

Educational relevance

Administrative routines

Sample University

Housing and leisure

Tuition fees and costs Pedagogical tools

Figure 13.1  Spider chart for external comparison.

254  Chapter Thirteen

Current performance

9

6 1

7

4

2

3

5 5 8

1

1

5

9

Importance

Figure 13.2  Performance matrix for establishing priorities.

Altogether, these analyses identified two related areas, ability to convey the academic message and pedagogical tools, as the areas to focus on. It was assumed that these were somehow connected to students’ ability to cooper‑ ate after completed studies. Focus was therefore put on the core process of teaching. The Department of Physics was selected as a pilot area. This was a field that could be difficult to teach, and this department had one of the low‑ est scores at the university and had experienced decreasing applications over the last few years. Results developed here would later be applied to the rest of Sample University.

13.3  Understanding the Current Processes and Performance Shortcoming After the Department of Physics was selected as the pilot area, a working group in the department was established to carry out the project. Together with the central pedagogical coordinator at the university, the working group consisted of: • Two professors from different academic fields • One research assistant • One person responsible for student laboratory work • Two students at different levels of study To enable the group to perform the work, substitute teachers were provided for the teaching personnel, and the participating students received an hourly wage.

Sample University—Improving Student Satisfaction  255

Introductory discussions concerning which processes could be defined within the department’s teaching activities led to viewing one course from start to finish as a process. As the different courses in the department were organized almost identi‑ cally, the group decided to start with only one course, assuming that any improve‑ ments achieved could be applied to the others. The first major task was to agree on how the process was currently conducted. Closer inspection revealed some minor differences between courses, so the work‑ ing group chose to document something close to a standard process. Through ple‑ nary discussions over a period of three to four weeks that focused on the level of detail to use, the group arrived at the flowchart shown in Figure 13.3.

Course planning

Welcoming the students

Course information

Preparations for lectures

Lecture

Calculation exercise

Summary lecture

Exam

Correcting and grading

Report on course grade

Figure 13.3  Flowchart for the course process.

256  Chapter Thirteen

Before the start of the semester, a brief course planning was performed where lecture topics were defined and other practical details were settled. At the beginning of the semester, the students taking the course were greeted with an introductory gathering where the course plan and curriculum were presented. The course was primarily composed of weekly cycles of theory lec‑ tures followed by a compulsory calculation exercise. The lectures were based on slide shows and blackboard teaching in relatively large classes of between 100 and 150 students. The calculation exercises were completed by the stu‑ dents on their own and then handed in for evaluation. Students were required to complete a certain number of exercises before they were allowed to take the final exam. Toward the end of the semester, a summarizing lecture was arranged, where the main contents of the curriculum were presented. The courses concluded with a written exam that was subsequently graded, and the grade was reported to the students. If a student failed the exam, the student had to retake it, up to three times. To enable measuring the improvement project’s results, the university man‑ agement measured the current performance level on the basis of its statistics on student grades. On a scale from 1 to 5, where 1 was weak and 5 was strong, the results for the department were as follows: • Usefulness of academic content: 3.4 • Ability to communicate the academic message: 2.3 • Pedagogical tools: 1.7 • Relevance for later jobs: 3.1 • Ability to cooperate: 1.9 Some central statistical information follows: • On a grading scale from 1.0 (worst) to 4.0 (best), the average grade in the department was 2.43 • The average grade varied somewhat between the student levels: the aver‑ age for ­first-­year students was 2.28, while the average of the ­highest-­level students was 3.11 • Within six months of graduation, 64 percent of the students were employed • After completing master’s degrees, 7 percent of the students proceeded to doctoral studies

Sample University—Improving Student Satisfaction  257

13.4  Analysis of the Performance Shortcoming To develop an overview of the pedagogical arrangement and the inability to get the message across to the students, the working group used a fishbone chart. Causes were sought for lacking ability to communicate the academic content. Through brainstorming and subsequent discussion and sorting of causes, the working group developed the diagram in Figure 13.4. Some of the main causes of the problems were the following: • The teaching was too theory focused, with little emphasis on application and industrial use • Coherency was missing in the studies and between individual courses • Learning was based on ­old-­fashioned principles (for example, one professor per 120 students, pure calculation exercises, and “learn and forget” exams) To better understand the interplay between the causes and their effects on the problem, the working group constructed a matrix diagram to analyze the rela‑ tionships between them. As there were really only two dimensions to the vari‑ ables—causes and problems—an L-matrix was used, but with the inclusion of a roof matrix to analyze connections between the different causes. The result is shown in Figure 13.5.

Academic content Too theory based

Study arrangement Lacking demonstration of applications

Missing course coherency Little industry focus

Too static Not updated about new research

No industry contact

Low progression between courses

Material too theory based Old textbooks

Individual learning

Calculation exercises too theory based

Only books No industry standards Teaching material

Old-fashioned lab equipment

No group work

Only lectures

Lacking communication ability

Exams are “learn and forget”

Teaching methods

Figure 13.4  Fishbone chart for lacking communication ability.

Too theory based

Lacking applications

Not updated on research

Lacking study coherency

Only lectures

Strong/ strong positive

Poor exam

Medium/ weak positive

Individual learning

Weak

Old teaching material

258  Chapter Thirteen

Academic content Ability to convey Pedagogical tools Relevance for later job Ability to cooperate Grade levels

Figure 13.5  Matrix diagram for Sample University.

Completing these analyses gave the working group good insight into the underlying causes for the disappointing performance levels. Thus, it was con‑ cluded that the problem analysis was complete and the search for solutions was started.

13.5  Generating Ideas and Improvement Proposals The working group envisioned using several improvement tools in this project. From the outset, many group members had ideas about what could be done dif‑ ferently and better; however, few had been clearly formulated and they were far from implementation. This richness of potential ideas led to the assumption that idealizing could be a suitable tool. Using it, the working group could brainstorm the ideal process for academic communication based on the group’s latent ideas and then use the results as a basis for improvement. The idealizing approach was selected, and elements of the ideal process were generated through brainwriting. Since the steps of the process were quite indepen‑ dent of each other, ideas could be generated for new ways of doing things at the individual steps. After a ­two-­day gathering where ideas were written and subse‑ quently sorted, grouped, and reformulated, the group ended up with the following improvement suggestions:

Sample University—Improving Student Satisfaction  259

• Start each course with a ­three-­hour motivational gathering where, in coop‑ eration between the lecturer and the students, a common understanding of the course objectives and content would be created • Communicate theory through colloquial sessions between the lecturer and a maximum of 15 students • Teach practical applications of theory by solving practical group exercise problems that are presented orally, and subsequently discuss solutions • Use one exam consisting of one individual theory part, oral or written, and one practical group work part with common grading • Encourage students to give feedback on arrangement and content through‑ out the course To supplement these proposals with even more suggestions, the working group wanted to perform a process benchmarking of a few other schools. Due to the lim‑ ited time and money for such an activity, this was reduced to a comparison against one other university. The state where Sample University is located presents an award each year to the higher education institution that has been noticed for excep‑ tional pedagogic merits. This year’s award was given to a small university special‑ izing in industrial management that is located in a neighboring town. A request to this college was favorably answered, and Sample University was welcomed to discuss its teaching methods. A ­half-­day visit with presentations and discussion around a typical course gave the following additional improvement suggestions: • Within each department, a complete evaluation of the course composition was conducted every third year. During this evaluation, the following issues were focused on: — All courses should fit together as a whole — The contents of the individual courses should not overlap too much — New research at the college itself and elsewhere should be included in the courses • In addition to colloquial teaching, and in line with what Sample University had already discussed, this college also used a system where the students, in groups and in turn, had to recite the most important aspects of last week’s chapter to the rest of the group. • In courses where this was possible, at the start of the semester, students would visit an industrial company or a research institution, where applica‑ tions of the course’s theory were demonstrated.

260  Chapter Thirteen

• Throughout the course, students worked on solutions to the problems observed during the excursion. At the end of the course, another visit to the company/research institution was arranged where the host organization presented its own solutions to these problems. • Before the exam, a ­four-­hour gathering was arranged to discuss the curricu‑ lum. The exam, however, was organized according to the suggestion from Sample University about a group exam. Combining Sample University’s own suggestions produced through brainwriting/ idealizing with the benchmarking outcomes created an extensive list of improve‑ ment proposals.

13.6  Implementation of Improvements The preceding phase, with the resulting set of improvement proposals, was concluded in early October. Objectives for the ­short-­term implementation at the Department of Physics were defined as planning, preparing, and arranging for a new educational system for the spring semester in three of the department’s six courses. The individual lecturers responsible for the courses, together with the working group, were responsible for this task. Because these lecturers did not teach any fall semester courses, they could devote the next three months to the implementation. To plan the work, a very simple implementation plan outlining some milestones was devised. Over the next three months, most of the following tasks were under‑ taken in all three courses: • Identifying new and better textbooks, as well as other multimedia materials • Outlining a program for colloquial teaching with smaller classes and stu‑ dent recitation • Planning industry and/or research excursions during the semester • Producing new group exercises demonstrating practical applications of the theory of the course • Developing a new exam format based on group solving • Developing a system for active feedback throughout the semester from the students In addition, a gathering was planned for the middle of October, where all the lec‑ turers would debate and coordinate the coherency of the set of courses offered by the department.

Sample University—Improving Student Satisfaction  261

Some resistance to the upcoming adjustments was expected from the faculty group, the administration, and the students. In an effort to ease the implementation and the changes it would entail, force field analysis was used. Following a brain‑ storming of forces for and against, the force field diagram shown in Figure 13.6 was constructed. Some countermeasures to these forces were identified as follows: • New information about the advantages of the new arrangement was distrib‑ uted to the students • Motivational gatherings where a pedagogical expert would describe experi‑ ences from other teaching institutions were organized for the lecturers • The rules were changed to facilitate common grading Prior to the course commencement in January, practically all of these changes were implemented. The only thing missing was the preparation necessary for a common exam, but there was plenty of time to do this over the course of the semester.

Changing the teaching arrangement

Forces for

Better study value

More rewarding study situation

Better grades

Easier to get a job

Increased number of applicants

Better university economy

Figure 13.6  Force field analysis for the implementation.

Forces against

Fear of unknown arrangement

Increased workload for the lecturers

Formal problems connected to common grading

Need for additional small colloquial rooms

262  Chapter Thirteen

After the spring semester courses were completed and in the months follow‑ ing, new measurements of the performance levels were conducted. The average results for the three courses were as follows: • Academic content usefulness: 3.9 • Communication ability: 4.5 • Pedagogical tools: 4.4 • Relevance for later jobs (few measures): 4.2 • Cooperation ability (few measures): 3.8 • Average grades: 3.28 As a result of the unconditional success, the department management decided to do the same in all the other courses. At the same time, the university management wanted the working group to make preparations for and support the same changes in all other departments at Sample University.

14 Green Carpet Seed—Revamping the Business Model

T

he Green Carpet Seed case study, to some extent, is also based on a ­real-­life improvement case. However, as was the case with the Sample University study, this project used only some of the tools presented in the book. There‑ fore, I have again taken the liberty of embellishing the story here and there, both to make it fit the overall improvement model of this book and to better illustrate the use of some tools and techniques. The case is from a company that produces seed for grass lawns and that found itself in trouble when sales continually declined for three years. Facing either extinction or a radical rethinking of its business, the company embarked on an ambitious improvement project that redefined its entire business model.

14.1  Description of the Organization Green Carpet Seed is a fairly small player in the seed industry, employing about 150 people. Established in 1952, the company is some 50 years old and remains family owned. It is based in a small town but is only 45 minutes from a large city. Its production facilities and administration offices are located in the same spot where the company was founded by the grandfather of the cur‑ rent CEO. Green Carpet Seed purchases seeds from local farmers, processes and treats them, mixes different types of seeds into combination blends, pack‑ ages the seeds, and sells them through garden centers, plant stores, and other outlets. The grass seed industry has radically changed since the founding of the company. Where there used to be small seed suppliers all over the country, there are now just a few left. This is a result of acquisitions of smaller units and the closing of many small production sites in favor of larger and more industrialized plants. Green Carpet Seed has avoided a takeover by responding

263

264  Chapter Fourteen

to tougher competition through rationalization efforts and automating parts of the production. For three years, however, sales have steadily declined, dropping 35 per‑ cent since their peak. Although the company has always enjoyed a comfort‑ able financial position, funds are eroding and there have been serious talks about whether it will make it through another year of falling sales. The CEO and his small executive group decided something drastic had to be done and launched an improvement project to look into how Green Carpet Seed could turn the situation around. A task force was established, bringing on board the sales manager, a biologist from R&D, a production worker, and a student from the local community college who was working on her master’s thesis and had proved extremely creative. They were charged with a mandate to propose ways to regain the lost sales. The project has now been completed, with a completely new business model as the result.

14.2  Development of Performance Priorities Having followed a strategy of remaining independent and not trying to grow by buying competitors, Green Carpet Seed had for many years been some‑ what cut off from discussions about the industry structure and its development. People had eyes and ears and saw the trends, but felt a little out of touch with the undercurrents of the sector. To reacquaint themselves with the latest think‑ ing and to gain a better insight into the competitive position of the company itself, the team wanted to conduct an analysis of the competitive situation in the industry. Together with a consultant who had worked with many players in various stages of the seed supply chain, the team spent a whole day discussing competitive forces. A long list of issues came up, some of them representing very scary sce‑ narios. Trying to eliminate items of less importance and focus on forces that would have a strong impact on the company, they ended up with a handful of issues they needed to be acutely aware of throughout the project: • The consolidation of seed suppliers was continuing, with a few companies growing quite large and still looking for smaller units to buy, either for their capacity and customers or to close them down as competitors. • Larger competitors courted local seed suppliers, and many had signed supply contracts with these competitors as well. In some cases, they were forced to sign exclusive contracts or preferred customer clauses that made it difficult to deliver to Green Carpet Seed.

Green Carpet Seed—Revamping the Business Model  265

• Some of the larger garden center chains developed their own brands for products such as seeds, flower bulbs, fertilizer, and even garden equipment. They still needed to cooperate with seed producers, but they negotiated very low prices in return for stable and larger volumes, and they would not allow the suppliers to market their own brands. • With home owners busier than ever, a general trend the past few years had been the search for ­easy-­to-maintain gardens. Some did away with lawns altogether in favor of patios and graveled areas; others kept their lawns but did no upkeep, meaning they reseeded less frequently. • Substitute products had existed for quite a while, but they were now being developed into stronger contenders to seeded lawns. ­Roll-­out lawn was easy to install and had proved more resistant to wear than before, and many home owners chose to replace the lawn every few years with new rolls instead of using seeds. Artificial grass was also becoming more common in private gardens and could further threaten seed sales. With all these threats looming on the horizon, the team felt anything but optimis‑ tic. Some of the team members had more than 20 years’ experience in the seed business, but none of them were really able to tell what would dictate the win‑ ners in the future. At one stage, they argued that lower prices would be the key to increasing sales, but this was followed by a conviction that customers did not mind the higher prices as long as the seeds gave them beautiful lawns. Realizing they were not able to debate their way to an answer, the sales manager suggested they try “calculating” an answer. In sales, they used various methods to determine the importance of sales channels and regions, and he figured a similar approach could be used here. Looking into this further, he realized that what he was looking for was the tool called criteria testing. The team members came up with a handful of critical suc‑ cess factors (CSFs) they believed determined the sales. They weighted them and listed business processes they believed influenced the CSFs. Carefully discuss‑ ing how each business process had an impact on each of the CSFs, they assigned impact factors and ended up with the matrix shown in Figure 14.1. To their surprise, a process they had seriously considered leaving out of the analysis came out on top: dealer management and support. During the analysis, they had made a point of not looking at the matrix itself, but assigning impact factors one by one to avoid letting their expectations of an answer dictate their assessments. When they finally compiled the numbers, this was a real ­eye-­opener. Checking the analysis again, they conceded that it made sense; the process truly did influence most of the success factors. With this result, the question was, what could be done about this process to gain a competitive edge?

Low prices

Solid packaging

Brand name

Convenience of use

3

1

2

1

3

Seed sourcing and supplier management

9

2

2

1

3

17

Seed processing

9

2

2

1

3

17

Product and process R&D

6

2

2

1

6

17

Seed packaging

3

1

6

1

6

17

Marketing and brand name promotion

3

1

2

3

3

12

Dealer management and support

6

3

2

3

6

20

Distribution

6

2

4

1

3

16

CSF Business processes

Weight

Total score

High-quality seeds

266  Chapter Fourteen

Figure 14.1  Criteria testing matrix for Green Carpet Seed.

14.3  Understanding the Current Processes and Performance Shortcoming The sales manager knew very well how the company worked with its distribution channels, but the rest of the team had virtually no insight into this part of the busi‑ ness. To educate them, the sales manager enlisted the help of a couple of sales‑ people in constructing a process model. Spanning many different departments and units in the total supply chain, a ­cross-­functional flowchart was the obvious choice. Somewhat simplified, the process is depicted in Figure 14.2. The sales manager was quite pleased with the flowchart, especially since one of the salespeople mentioned that the neighborhood kids who mowed lawns were affected by the trouble the seeds from Green Carpet Seed caused in terms of good grass growth. They added this “player” to the chart and pre‑ sented it to the rest of the team. Having looked at it, the others asked what the difference was between this process and distribution. The first reflex of the sales manager was to dismiss the question as silly: they were two completely different things. Looking at the chart again, however, he realized the question was warranted. Thinking they were modeling the dealer management and support process, they had indeed drawn the sales and distribution process. Mulling this over, the sales department people had to acknowledge that they didn’t really have much of a dealer support process. They accepted their orders well enough and sometimes went to see the dealers, generating more orders. However, once the orders were

Green Carpet Seed—Revamping the Business Model  267

Soliciting/ receiving incoming orders

Warehouse

Sales

Dealer management and support

Truck transport

Purchase decision

Independent stores

Chain stores

Chain purchasing

Transport company

Order packaging

Receipt of seeds

Purchase decision

Receipt of seeds

Purchase of seeds

Lawn mower

Customer

Shipment order

Sowing

Lawn mowing

Figure 14.2  ­Cross-­functional flowchart for the dealer management and support process.

secured, everything was handed over to someone else to deliver the seeds to the stores. There was very little cooperation with the sales outlets about things like promotions, campaigns, meeting customers, and so on. Another realization came when the student on the team asked questions about the category of the kids who mowed lawns, added just for fun. Why do garden own‑ ers pay kids to mow their lawn? And if they hire mowers, do they also hire ­gardeners?

268  Chapter Fourteen

If the garden owners don’t tend to their lawns themselves, how do they know if and when they need seeding? And if they don’t do it, who does? The kids mowing the lawns? The gardeners? She seemed to have an idea behind these questions, and although the sales manager didn’t quite see where she was headed, something stirred at the back of his mind. He knew that there were many successful companies in the city that provided only one service: tending to plants in office buildings, train stations, and other public spaces. Sometimes the owners of the building or facility owned the plants; in other cases, these companies owned the plants and leased them to customers, in addition to providing care services for them. Both the sales manager and the student seemed to arrive at the same basic idea at almost the same time: was it possible to add a service element to the sale of seeds? As with the plant service companies, this could be a ­revenue-­generating service of its own and at the same time help sell more of the company’s seeds. The rest of the team was quite enthusiastic about the idea but had many objections as well. To move ahead and decide whether this could be a solution, they debated the idea thoroughly.

14.4  Generating Improvement Ideas Although the team members believed they might be on to something with this added service idea, they felt a need to view it from several angles. The biologist had learned the technique of the six thinking hats at her university and suggested they try the approach. Getting the others to agree, she quickly explained the prin‑ ciples to them and they met one morning, armed with plenty of coffee, to run a session. Two more people joined them: one of the salespeople who had helped out with the flowchart and the CEO himself. They shared the hats among them, even putting on real hats, and got ready to discuss how they could add a service element to the seed sales. The session ran for almost four hours, stopping only for a short break. When they finished and the stu‑ dent, wearing the blue hat, summarized their debate, they had several very creative ideas on the table: • Closer cooperation with dealers. Working more actively with dealers, they could train them in helping customers select the right seeds and determine the correct time to use them. They could also train them in arranging cam‑ paigns in their stores. Launching an annual lawn week, they could have stands in the stores, supported by advertisements in different media, to increase general lawn awareness. They could offer expert assistance to cus‑ tomers who came to the stores during that week, and sales would probably increase during that time, as well as the rest of the year, due to the knowl‑ edge gained by the customers.

Green Carpet Seed—Revamping the Business Model  269

• A lawn consulting service could be offered to customers. An expert would study the lawn, or the area where the customer wanted to establish a lawn, and give advice on which type of seed to use, when to sow, how to treat the lawn, and so on. • Lawn mowing and lawn tending services could be developed. Working with both students and other ­low-­cost employees, the company could offer lawn mowing, airing, reseeding, and other care services. • They could work with seed dealers to develop an automatic seed delivery process. The thinking was that a computer system could keep track of all the customers, where they lived, when the lawn was established, which seeds were used, soil conditions, weather conditions during the year, and so on. Using this information, the computer would calculate when reseed‑ ing was beneficial and would either remind the customers to come and buy seeds or deliver seeds directly to them. This service could be combined with tending services, where a lawn care person would sow the seeds and water afterward. • Going beyond the lawn, it would be possible to set up an entire garden care service system, where cooperation with gardeners, lawn tenders, snow removers, and so on, would provide a complete range of services to customers. Several of these ideas could be combined, and common to all was that they would directly or indirectly lead to increased sales, with many of the services also making money on their own.

14.5  Developing an Improved Process The team members, including the CEO, were extremely enthusiastic about their many ideas. Realizing that few, if any, of these approaches had been tried before in the industry, they felt they were on the brink of a true breakthrough. They understood that most of the ideas, if implemented, could be rather quickly copied, but they hoped the ­first-­mover advantage would keep them ahead of the competition for a while. Although they had many more or less radical ideas, they still didn’t quite see how such new services should be organized and designed. The next step was to attempt designing the services and the processes required to deliver them in some more detail. Understanding how much change this would require, they quickly adopted the term reengineering and set about drawing up the new processes. Reading up on BPR, they had no doubt that they would need to use clean sheet reengineering.

270  Chapter Fourteen

They brought in a person from one of the dealers they knew well and the founder of one of the ­better-­known plant service companies in the nearby city to strengthen the team. To guide the discussion, they went through many of the rec‑ ommended questions to prompt their creativity further: • Which needs of the customers can we fill with a new service? • With which organization can we offer services that fill these needs? Through the stores, directly from the company, or through an independent service company? • Where and when can the service be delivered? • How will the needs be fulfilled? What business processes, equipment, per‑ sonnel, and so on, will be required? • Can we learn something from the plant service industry and how it launched its services? In the end, they decided to proceed with two main ideas. The first was the annual lawn week event. This idea was fairly simple to develop further, and it was quite similar to campaigns run by other industries—for example, the annual “Go bike!” event that local sporting goods stores arranged every spring. A dedicated team from sales and marketing was given the task of pursuing this idea. The other idea was an integrated lawn service, which was a combination of several of the ideas they had developed during the six thinking hats session. Hav‑ ing discussed it for several weeks, including calculating cost and income, they concluded that such a service would probably be difficult to run at much of a profit. They had therefore agreed that it should be priced so low that it would attract a maximum number of customers. This way, it would be most effective in fulfilling its main purpose: generating higher sales of seeds, which would continue to be the main source of revenue. Without going into details, the service was intended to be organized and designed as follows: • Green Carpet Seed would set up a new company to operate the service. It had considered a joint venture with one of the large garden center chains, but feared “marrying” one would cause other dealers to switch to other suppliers. • The new company would deliver a ­three-­tier integrated lawn service pack‑ age consisting of: — The aforementioned lawn consulting service. This would ensure that new customers would be continually recruited into the program.

Green Carpet Seed—Revamping the Business Model  271

— The lawn tending service, employing a network of students from local high schools and colleges and overseen by a team of gardeners. — An automatic seed delivery service monitored by computer, as described earlier. • This computer system would very much be at the core of the ­full-­service package, managing the delivery of the various services. • Any fertilizer and equipment required would be sourced through local deal‑ ers, but Green Carpet Seed would be used for seed deliveries. The final proposal for the establishment of this new integrated service package was naturally much more extensive than this and outlined a number of details. Com‑ prehensive financial calculations supported the service design to show its feasibil‑ ity and impact on seed sales. Having delivered the proposal to the CEO, the team temporarily disbanded to await the decision of the executive team.

14.6  Implementation of the New Service The executive team failed to decide, being squarely split down the middle, and the proposal was sent to the board. Because Green Carpet Seed was a ­family-­owned company, the board tended to be conservative and was skeptical of the idea. Speak‑ ing warmly for the idea and warning that bankruptcy was the likely outcome if the plan was not implemented, the CEO managed to secure the board’s approval in the end. The improvement team was asked to resume its work and to start planning the establishment of the company. It searched for a manager of the company, recruited one from one of the plant service companies, and proceeded with the planning. In this phase, several versions of force field diagrams were developed, addressing dif‑ ferent stakeholders that would be affected by the plans. These diagrams gave the team many important insights into how competitors, dealers, and future employees could react to this move. Next, the team developed an overall implementation plan in the form of a tree diagram. This was combined with the outcome of the force field analyses to construct a PDPC for the process. A simplified version of this is shown in Figure 14.3. The new company was launched a little more than a year after the project was started. At this point, red flags were waving, signaling the imminent death of the mother company. Only by convincing the bank that the new service would revital‑ ize sales was Green Carpet Seed allowed to go on for another few months. Right from the beginning, customers flocked in to sign up for the service pack‑ age. Impressed with the responsibility the company would take for their gardens,

Working more closely with dealers

Appoint liaison person X

Finding good name

Hire marketing bureau O

Create Web appointment system X

Work with local schools and colleges X

Work with local employment agency O

Develop lawn tending

Organizing the temporary employees

Develop lawn consulting

Develop new services

Finding enough temporary employees

Recruit employees

Figure 14.3  PDPC for Green Carpet Seed.

Recruit manager

Legal paperwork

Establish new company

Establish clear contracts O

Organizing the dealer interface

Develop seed delivery

Launch new integrated lawn service

O

Hire consultant

Low in-house IT expertise

Specify system requirements

Search for global vendors X

Few vendors available

Request tenders

Install computer system

Develop training program O

Sufficient training for users

Select vendor and purchase system

272  Chapter Fourteen

Green Carpet Seed—Revamping the Business Model  273

they were happy to pay the modest fee asked. After six months of operation, the new company had a client list of close to 2000 lawn owners. The operating margin for the new service company was only 1.2 percent, barely breaking even. Seed sales for the parent company, on the other hand, had increased by 16 percent, almost catching up with half of the lost sales.

Appendix Template Package

T

his section contains different templates or forms for tools that can be used in your improvement process. I considered developing templates that could be used on your computer, but decided that many of these tools and techniques do not easily lend themselves to such electronic templates. Scaling, the number of fields, how things are drawn, and so on, make it very difficult to make templates that could accompany the book on a CD. Furthermore, software is available for many of the tools that are widely used. Even in programs such as Excel and Visio in the Microsoft Office suite, many of these tools appear as standard options. Thus, the templates included here are those that are suitable for presentation on paper. You can copy them directly from the book, preferably in an enlarged format, and write on them, or they can simply function as input for creating your own templates. Use of each template is explained in the chapter where the specific tool is described.

275

276  Appendix

SWOT Analysis Strengths

Weaknesses

Opportunities

Threats

Performance

Trend Analysis

Time

Today

Template Package  277

Spider Chart CSF1

CSF8

CSF2

CSF7

CSF3

CSF4

CSF6

CSF5

Performance Matrix

Current performance

9

5

1

1

5

Importance

9

278  Appendix

Criteria Testing CSF Processes

Weight

1

2

3

4

5

Total score

Template Package  279

Quality Function Deployment (QFD)

Direction of improvement Competitive benchmarking Importance

Hows

Whats 1

Targets Competitive benchmarking

5 4 3 2 1 Absolute importance Relative importance Organizational difficulty Absolute importance corrected for difficulty Relative importance corrected for difficulty Roof

Roof

Weight

Strong positive Positive Negative Strong negative

Strong Medium Weak

9 3 1

2

3

4

5

280  Appendix

Relationship Mapping Own company

Flow of goods

Information

Order

Discussion/ Negotiation

Cross-Functional Flowchart

Template Package  281

Check Sheet

Problem

Period 1

Period 2

Period 3

Total number of occurrences per problem

Total number of problems per period

Pareto Chart

Problems

100%

50%

A

B

C

D

E

F

282  Appendix

Cause-and-Effect Chart Employees

Methods

Environment

Equipment

Five Whys Analysis

Why? Why? Why? Why? Why?

Template Package  283

Scatter Chart

Histogram Frequency

Category

284  Appendix

Relations Diagram in out

in out

in out

in out

in out

in out

Is–Is Not Analysis

Performance shortcoming What occurs? What objects are affected? Where does the problem occur? When does the problem occur? Extent of shortcomings?

Who is involved?

Is

Is not

Distinction

Number of votes

F

E

D

C

B

A

A/B

A/C

A/D

A/E

A/F

B/C

B/D

B/E B/F

C/D

C/E

C/F

D/E D/F

E/F

Total

Template Package  285

Paired Comparisons

286  Appendix

Control Chart UCL =

Center line

LCL =

Force Field Analysis

Forces for

Forces against

Bibliography Ackerman, R. B., R. J. Coleman, E. Leger, and J. C. MacDorman. 1987. Process Quality Management & Improvement Guidelines. Indianapolis, IN: Publication Center, AT&T Laboratories. Akao, Yoji, ed. 1990. Quality Function Deployment: Integrating Customer Requirements into Product Design. Cambridge, MA: Productivity Press. American Productivity & Quality Center (APQC). 1993. Basics of Benchmarking (Course Material). Houston, TX: APQC. Andersen, Bjørn, and Tom Fagerhaug. 2001. Performance Measurement Explained: Designing and Implementing Your ­ State-­of-the-Art System. Milwaukee, WI: ASQ Quality Press. Andersen, Bjørn, and Bjørnar Henriksen. 2004. Performance Measurement System Suc‑ cess Depends on Design Alignment with the Core Purpose of Its Implementation. Proceedings from PMA 2004—Public and Private Conference, Performance Measure‑ ment Association, Edinburgh, Scotland, July 28–30. Andersen, Bjørn, and ­Per-­Gaute Pettersen. 1996. The Benchmarking Handbook: ­Step-­byStep Instructions. London: Chapman & Hall. Aune, Asbjørn. 1985. Kvalitetssirkler: Problemløsningsgrupper for personlig vekst, kvalitet og produktivitet [Quality Circles: Problem Solving Teams for Personal Growth, Qual‑ ity, and Productivity]. Oslo, Norway: Universitetsforlaget. ———. 1993. Kvalitetsstyrte bedrifter [Quality-Managed Companies]. Oslo, Norway: Ad Notam. Balm, Gerald J. 1992. Benchmarking: A Practitioner’s Guide for Becoming and Staying Best of the Best. Schaumburg, IL: QPMA Press. Bendell, Tony, Louise Boulter, and John Kelly. 1993. Benchmarking for Competitive Advantage. London: Pitman Publishing. Bertels, Thomas, ed. 2003. Rath & Strong’s Six Sigma Leadership Handbook. Hoboken, NJ: John Wiley & Sons. Bogan, Christopher E., and Michael J. English. 1994. Benchmarking for Best Practices: Winning through Innovative Adaptation. New York: ­McGraw-­Hill. Bossert, James L. 1991. Quality Function Deployment: A Practitioner’s Approach. Mil‑ waukee, WI: ASQC Quality Press. Boxwell, Robert J. 1994. Benchmarking for Competitive Advantage. New York: ­McGraw-­Hill. Bredrup, Harald. 1995. Performance Measurement in a Changing Competitive Industrial Environment: Breaking the Financial Paradigm. PhD thesis, Norwegian Institute of Technology, Trondheim, Norway.

287

288  Bibliography

Burr, J. T. 1990. “The Tools of Quality: Part I: Going with the Flow(chart).” ASQC Quality Progress, June, pp. 64–67. Camp, Robert C. 1989. Benchmarking: The Search for Industry Best Practices That Lead to Superior Performance. Milwaukee, WI: ASQC Quality Press. ———. 1994. Business Process Benchmarking: Finding and Implementing Best Practices. Milwaukee, WI: ASQC Quality Press. ———, ed. 1998. Global Cases in Benchmarking: Best Practices from Organizations around the World. Milwaukee, WI: ASQ Quality Press. Codling, Sylvia. 1992. Best Practice Benchmarking: The Management Guide to Successful Implementation. Dunstable, England: Industrial Newsletters. Cohen, Lou. 1995. Quality Function Deployment: How to Make QFD Work for You. Cam‑ bridge, MA: ­Addison-­Wesley. Collins, Jim. 2001. Good to Great: Why Some Companies Make the Leap and Others Don’t. New York: HarperBusiness. Conti, Tito. 1997. Organizational ­Self-­Assessment. London: Chapman & Hall. Davenport, T. H. 1993. Process Innovation: ­ Re-­engineering Work through Information Technology. Boston: Harvard Business School Press. Day, Ronald G. 1993. Quality Function Deployment: Linking a Company with Its Custom‑ ers. Milwaukee, WI: ASQC Quality Press. De Bono, Edward. 1985. Six Thinking Hats. Boston: Little, Brown, and Company. De Feo, Joseph A., and William W. Barnard. 2004. Juran Institute’s Six Sigma Break‑ through and Beyond. New York: ­McGraw-­Hill. Deming, W. E. 1986. Out of the Crisis: Quality, Productivity and Competitive Position. Cambridge, MA: Cambridge University Press. Dertouzos, M. L., R. K. Lester, R. M. Solow, and the MIT Commission on Industrial Pro‑ ductivity. 1989. Made in America: Regaining the Productive Edge. Cambridge, MA: MIT Press. D’Herbemont, O., and B. Cesar. 1998. Managing Sensitive Projects: Lateral Approach. New York: Routledge. Eastman Kodak Company. 1990. Quality Leadership Process Guidebook. Rochester, NY: Eastman Kodak Company. ENAPS. 1997. Deliverable F3-4: A Set of Refined and Agreed Performance Indicators Defined by Business Processes. Galway, Ireland: ENAPS. Ericsson Quality Institute. 1993. Business Process Management. Gothenburg, Sweden: Ericsson. Fellers, G. 1992. The Deming Vision: SPC/TQM for Administrators. Milwaukee, WI: ASQC Quality Press. Galloway, Dianne. 1994. Mapping Work Processes. Milwaukee, WI: ASQC Quality Press. Gitlow, Howard, Alan Oppenheim, and Rosa Oppenheim. 1995. Quality Management: Tools and Methods for Improvement. Burr Ridge, IL: Irwin. Goldratt, Eliyahu M. 1990. What Is This Thing Called Theory of Constraints and How Should It Be Implemented? Great Barrington, MA: North River Press. Goldratt, Eliyahu M., and Jeff Cox. 1992. The Goal: A Process of Ongoing Improvement. 2nd ed. Great Barrington, MA: North River Press.

Bibliography  289

Govers, C.P.M. 1994. What and How of Quality Function Deployment (QFD). Proceedings from the Eighth International Working Seminar on Production Economics. Igls/Inns‑ bruck, Austria, February 21–25. Griffiths, John. 1993. “Driving Out the Old Regime.” Financial Times, August, p. 20. Hammer, M., and J. Champy. 1993. Re-engineering the Corporation: A Manifesto for Busi‑ ness Revolution. New York: Harper Business. Harrington, H. James. 1991. Business Process Improvement: The Breakthrough Strategy for Total Quality, Productivity, and Competitiveness. New York: ­McGraw-­Hill. Hayes, R. H., and S. C. Wheelwright. 1984. Restoring Our Competitive Edge: Competing through Manufacturing. Boston: John Wiley & Sons. Hronec, Steven M. 1993. Vital Signs: Using Quality, Time, and Cost Performance Mea‑ surements to Chart Your Company’s Future. New York: Amacom. Hunt, Daniel. 1996. Process Mapping: How to Reengineer Your Business Processes. New York: Wiley. Hutchins, David. 1985. Quality Circles Handbook. New York: Nichols. Imai, M. 1986. KAIZEN: The Key to Japan’s Competitive Success. New York: Random House Business Division. Johnson, H. T., and R. S. Kaplan. 1987. Relevance Lost: The Rise and Fall of Management Accounting. Boston: Harvard Business School Press. Kano, N., N. Seraku, F. Takahashi, and S. Tsuji. 1984. “Attractive Quality and ­Must-­Be Quality.” The Journal of the Japanese Society for Quality Control, April, pp. 39–48. Kaplan, R. S., ed. 1990. Measures for Manufacturing Excellence. Boston: Harvard Busi‑ ness School Press. Kaplan, R. S., and D. P. Norton. 1996. The Balanced Scorecard. Boston: Harvard Business School Press. ———. 2004. Strategy Maps: Converting Intangible Assets into Tangible Outcomes. Bos‑ ton: Harvard Business School Press. Kubeck, Lynn C. 1995. Techniques for Business Process Redesign: Tying It All Together. New York: John Wiley & Sons. Kume, H. 1992. Statistical Methods for Quality Improvement. Tokyo, Japan: Association for Overseas Technical Scholarships (AOTS). Lawlor, Alan. 1985. Productivity Improvement Manual. Aldershot, England: Gower Publishing. Lynch, R. L., and K. C. Cross. 1991. Measure Up! Yardstick for Continuous Improvement. Cambridge, MA: Blackwell Business. McCormack, Kevin. 2001. “Business Process Orientation: Do You Have It?” ASQ Quality Progress, January, pp. 51–57. McNair, Carol J., and Kathleen H. J. Leibfried. 1992. Benchmarking: A Tool for Continu‑ ous Improvement. New York: Harper Business. Meisel, Robert M., Steven J. Babb, Steven F. Marsh, and James P. Schlichting. 2007. The Executive Guide to Understanding and Implementing Lean Six Sigma. Milwaukee, WI: ASQ Quality Press. Miller, K. 2002. The Change Agent’s Guide to Radical Improvement. Milwaukee, WI: ASQ Quality Press. Mintzberg, H. 1994. The Rise and Fall of Strategic Planning. New York: The Free Press.

290  Bibliography

Mizuno, Shigeru, ed. 1988. Management for Quality Improvement: The 7 New QC Tools. Cambridge, MA: Productivity Press. Peppard, Joe. 1998. “Benchmarking Business Process: A Framework and Classification Scheme,” in Proceedings of the ESPRIT—COPERNICUS Symposium Distributed Enterprise, Intelligent Automation, and Industrial Benchmarking, Wroclaw, Poland. Peppard, Joe, and Philip Rowland. 1995. The Essence of Business Process ­Re-­engineering. Hemel Hempstead, England: Prentice Hall. Porter, M. E. 1980. Competitive Strategy: Techniques for Analyzing Industries and Com‑ petitors. New York: Free Press. ———. 1985. Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press. ———. 1990. Competitive Advantage of Nations. London: The MacMillan Press. Przekop, Penelope. 2006. Six Sigma for Business Excellence: A Manager’s Guide to Super‑ vising Six Sigma Projects. Milwaukee, WI: ASQ Quality Press. Robson, G. D. 1991. Continuous Process Improvement: Simplifying Work Flow Systems. New York: The Free Press. Robson, Mike. 1988. Quality Circles: A Practical Guide. Aldershot, England: Gower. Rolstadås, Asbjørn, ed. 1995. Performance Management: A Business Process Benchmark‑ ing Approach. London: Chapman & Hall. Savage, Grant T., Timothy W. Nix, Charlton J. Whitehead, and John D. Blair. 1991. “Strate‑ gies for Assessing and Managing Organizational Stakeholders.” Academy of Manage‑ ment Executive 5 (2): 61–75. Scholtes, Peter R., Brian L. Joiner, and Barbara J. Streibel. 2003. The Team Handbook, 3rd ed. Madison, WI: Oriel Incorporated. Sink, D. S., and T. C. Tuttle. 1989. Planning and Measurement in Your Organization of the Future. Norcross, VA: Industrial Engineering and Management Press. Spendolini, Michael J. 1992. The Benchmarking Book. New York: Amacom. Spitzer, Dean R. 2007. Transforming Performance Measurement: Rethinking the Way We Measure and Drive Success. New York: Amacom. Stalk, George Jr., and Thomas M. Hout. 1990. Competing against Time: How ­Time-­Based Competition Is Reshaping Global Markets. New York: The Free Press. Straker, David. 1995. A Toolbook for Quality Improvement and Problem Solving. London: ­Prentice-­Hall. Swanson, Roger C. 1995. The Quality Improvement Handbook: Team Guide to Tools and Techniques. London: Kogan Page. Syan, Chanan S., and Unny Menon, eds. 1994. Concurrent Engineering: Concepts, Imple‑ mentation and Practice. London: Chapman & Hall. Watson, Gregory H. 1992. The Benchmarking Workbook: Adapting Best Practices for Per‑ formance Improvement. Cambridge, MA: Productivity Press. ———. 1993. Strategic Benchmarking: How to Rate Your Company’s Performance against the World’s Best. New York: John Wiley & Sons. Wikipedia. 2006. Stakeholder. http://en.wikipedia.org/wiki/Stakeholder. Zairi, M., and Paul Leonard. 1994. Practical Benchmarking: A Complete Guide. London: Chapman & Hall.

Index observation, 231–232 planning, 227–228 search, 228–230 process of, 225–226, 226f benchmarking wheel, 227, 228f Black Belts, 206–207 bottleneck analysis, 151–155 example of, 153–154, 155f identifying bottlenecks in, 153 bottleneck resources, 151–152, 152f BPR. See business process reengineering (BPR) brainstorming, 157–159 approaches to, 158 benefits of, 158 for clean sheets, 215, 215f example of, 159 for identifying stakeholders, 11 procedure for, 158–159 structured, 158 unstructured, 158 brainwriting, 159–160 methods of, 160 procedure for, 160 bureaucracy, eliminating, 168–169 business processes. See also modeling, of business processes; processes classifying, 34–36 defined, 32–34 with department boundaries, 30, 30f departments and, 27–32 derived from strategy and stakeholders, 37, 37f documenting, 39 generic, 29f hierarchy of models for, 40–43, 41–42f horizontal, 28–32 identifying organizational, 37–39 modeling, 38–43

Page numbers followed by f or t refer to figures or tables, respectively.

A ADT analysis, 239–241 defined, 239 example of, 240–241 procedures for, 240 arithmetic mean value, 183 attribute data chart types for c, 188 np, 188 p, 188 u, 188 automation, 213–214 80-20 rule, 214 Automotive Industry Action Group (AIAG), 35

B bar charts. See histograms bell-shaped curves, 184, 185f benchmarking, 221–236 benefits of, 223 combinations of types of, 225, 225f competitive, 223 defined, 221–226, 222f generic, 223 internal, 223 strategic, 225 benchmarking partners, 229–230, 230f benchmarking studies conducting, 227–233 examples of, 233–236 models for, 226–227, 226f phases of analysis, 232–233 implementation, 233

291

292  Index business process improvement. See also improvement project planning; tools defining organizational strategy for, 17–18 framework for, 4–7, 5f model of, 108–109 organizational modes for, 95–100 performance measurement in, 65–67 planning projects for, 75–76 skills for, 93–94 business process reengineering (BPR), 207–221 approaches to, 208–209 case study of, 269–271 defined, 208 example of, 217–221 implementation phase, 216–217 process for, 209–217, 210f implementation, 210 planning, 210–211 reengineering, 210, 211–215 transformation, 210, 215–216 types of, 209

C card method, for brainwriting, 160 case studies Green Carpet Seed, 263–273 Sample University, 251–262 cause-and-effect charts, 128–131 example of, 131 fishbone charts, 128–130 process charts, 131 template, 282 c charts, for attribute data, 188 constructing, 196 change process, formula for, 245–246 check sheets, 117–119, 136f example of, 118–119, 118f procedure for, 117–118 template, 281 chronic variation, 183, 185 clean sheet reengineering, 209, 214–215 cluster sampling, 114 competitive force analysis, 20–21 example of, 21–22 constraints, theory of, 151, 152 continuous improvement, need for, 4 control charts. See also statistical process control (SPC) for attribute data c, 188 np, 188 p, 188 u, 188

constructing, 189–196 c charts, 196 np charts, 194–195 p charts, 193–194 u charts, 195–196 x and R charts, 191–193 X bar and R charts, 189–191, 191f, 192f interpreting, 196–200 template, 286 types of, 186–188 control limits, 186–187 CPM (Critical Path Method), 242 Crawford slip method, 159–161 criteria testing, 83–85 case study of, 265, 266f example of, 85 form for, 84f procedure for using, 84 template, 278 critical incident technique, 123–125 example of, 124–125 process of, 124 critical success factor (CSF) case study of, 265 defined, 83 cross-functional flowcharts, 51–54 case study of, 266, 267f constructing, 52–53 example of, 52f, 53–54 size of, 103 swim lanes in, 51, 52f template, 280 cross-functional teams, 102–103 defined, 103 work rules for, 103 CSF. See critical success factor (CSF) C-shaped matrix diagrams, 146 cycle times, reducing, 172–174

D data attribute, 186–188 variable, 186–188 de Bono, Edward, 161–162 defective vs. defects, 187–188 Deming, W. Edwards, 6 Deming wheel, 6–7, 7f department boundaries, 30, 30f departments, business processes and, 27–32 development processes, 36 DMAIC (define, measure, analyze, improve, and control) process, of Six Sigma, 204–206 documentation of business processes, 39

Index  293

E

I

80–20 rule, 125–126 automation and, 214 e-learning approaches, 94 emergent strategies, 18 ENAPS (European Network or Advanced Performance Studies) project, 36, 36f ESIA (elimination, simplification, integration automation) rules, 212–213 European Foundation for Quality Management (EFQM), 35

idealizing, 177–178, 177f example of, 178 IDEF0, 48–51, 50f implementation phase, of BPR, 210, 216–217 improvement culture, stimulating, 103–105 improvement project planning overview of, 75–76 tools for criteria testing, 83–85 performance matrix, 81–83 quality function deployment (QFD), 85–92 spider charts, 78–81 trend analysis, 76–78 improvement projects preconditions for, 94 training for, 94 improvements maintenance for, 3, 3f necessity for, 1–2 stakeholder analysis for, 10 stakeholders expectations of, 9–17 tools for benchmarking, 221–236 business process reengineering, 207–221 idealizing, 177–178 quality function deployment, 179–182 Six Sigma, 204–207 statistical process control/control charts, 182–203 streamlining, 167–176 tools for implementing, 237–239 ADT analysis, 239–241 force field analysis, 245–249 tree diagrams, 241–243 improvement teams, 254–255 composing, 98 example of, 100 requirements for selecting members to, 99–100 roles of, 98–99 indicators, for measurement financial vs. nonfinancial, 71–72 hard vs. soft, 70–71 leading vs. lagging, 72–74 internal benchmarking, 223 competitive, 223 generic, 223 performance, 224 process, 224 International Benchmarking Clearinghouse (IBC), 35 interviews, 115–117

F feedback from customers, 9 from employees, 9 financial indicators, 71–72, 72t fishbone charts, 128–130, 128f, 130f, 257f producing, 129 five whys analysis, 132–133, 132f example of, 133 procedure for, 132 template, 282 flowcharts, 255f cross-functional, 51–54 designing, 48 divided into process areas, 54–55 example of, 47f, 48, 49f paper/pencil or computers for constructing, 62–63 several-leveled, 55–58 with statistics, 58–62, 61f symbols for, 46–47 traditional, 46–51 force field analysis, 245–249, 261, 261f diagram for, 247f example of, 248–249 procedure for using, 247 template, 286

G gallery method, for brainwriting, 160 gap analysis, 232–233 Green Belts, 207 Green Carpet Seed, case study of, 263–273

H hard indicators, 70–71, 70t histograms, 136–142, 136f constructing, 136–138 patterns indicating problems in, 139–142 template, 283

294  Index is–is not analysis, 148–151 example of, 150–151 steps for undertaking, 14 template, 283

K Kano model, 12–14, 13f, 16

L lagging indicators, 72–74 leading indicators, 72–74 L-shaped matrix diagrams, 146, 146f

M mailboxes, for feedback, 9 Malcolm Baldrige National Quality Award, 204 Master Black Belts, 206–207 matrix diagrams, 145–148, 258f C-shaped, 146 example of, 148 L-shaped, 146, 146f relations symbols for, 147, 147f roof-shaped, 146, 146f steps for using, 147 T-shaped, 146, 146f X-shaped, 146, 146f Y-shaped, 146, 146f matrix organizations, 98 modeling, of business processes, 38–43 approaches for cross-functional flowcharts, 51–54 flowcharts divided into process areas, 54–55 flowcharts with statistics, 58–62, 61f relationship mapping, 43–46 several-leveled flowcharts, 55–58 traditional flowcharts, 46–51 Motorola, 204

N nominal group technique (NGT), 163–165 example of, 164–165 procedure for, 163–164 non-bottleneck resources, 151 nonfinancial indicators, 71–72 non-value-adding activities (NVA), 170–172 normal distribution curves, 184, 185f normalizing, defined, 232 np charts, for attribute data, 188 constructing, 194–195

O optimized production technology (OPT) theory, 151 organizational value-adding activities (OVA), 170–172 organizations stimulating improvement culture of, 103–105

P paired comparisons, 165–166 example of, 166 procedure for, 165–166 template, 285 Pareto, Vilfredo, 125 Pareto charts, 125–127, 126f example of, 127 steps for using, 126–127 template, 281 p charts, for attribute data, 188 constructing, 193–194 PDPC. See process decision program chart (PDPC) performance benchmarking, 224 performance matrix, 254f defined, 81 example of, 82–83 quadrants of, 81, 82f template, 277 using, 82 performance measurement in business process improvement, 65–67 implementing, 67–68, 69f indicators for, 70–74 financial vs. nonfinancial, 71–72, 72t hard vs. soft, 70–71, 70t leading vs. lagging, 72–74 PERT (Program Evaluation and Review Technique), 242 planning phase, of BPR, 210–211 Porter’s value chain, 35 primary processes, 35–36, 35f problem concentration diagrams, 119–122 example of, 120–122 steps for using, 120 process areas, flowcharts divided into, 54–55 process benchmarking, 224 process capability, 200–202 process cycle times, reducing, 172–174 process decision program chart (PDPC), 243–245, 244f approach for using, 243–244 case study of, 271, 272f example of, 245 process departments, 97 processes. See also business processes development, 36 primary, 35–36, 35f support, 35, 35f process ownership, 95

Q QFD. See quality function deployment (QFD) qualitative relations diagrams, 142–143, 142f

Index  295 quality circles, 9, 100–102 defined, 101 objectives of, 101 organization of, 101, 102f shortcomings of, in the West, 101–102 quality function deployment (QFD), 85–92, 179–182 example of, 90–92, 180–181 process for conducting, 86–88 structure of, 86, 86f template, 279 quantitative relations diagrams, 143–145, 144f questionnaires, 115–117, 252

R random sampling, 114 R charts. See X bar and R charts, constructing real value-adding activities (RVA), 170–172 redundancy, eliminating, 169 reengineering phase, of BPR, 211–215 clean sheet approach, 214–215 systematic reengineering approach, 212–214 relations diagrams, 142–145 example of, 143, 145 qualitative, 142–143, 142f quantitative, 143–145, 144f template, 283 relationship mapping, 43–45, 44f example of, 45–46 template, 280 relations symbols, for matrix diagrams, 147, 147f resources bottleneck, 151 non-bottleneck, 151 roof-shaped matrix diagrams, 146, 146f

S SADT. See Structured Analysis and Design Technique (SADT) sample size, deciding on, 114 Sample University, case study of, 251–262 sampling, 113–115 cluster, 114 example of, 115 issues for, 114 random, 114 stratified, 114 systematic, 114 scatter charts, 134–136, 134f example of, 135–136 steps for using, 135 template, 283 several-leveled flowcharts, 55–58 creating, 56–57 example of, 58

Six Sigma, 204–207 DMAIC process for, 204–206 organizational structure for, 206–207 six thinking hats, 161–163 case study of, 268–269 colors for, 162 example of, 163 procedure for, 162 skills, for business process improvement, 93–94 soft indicators, 70–71, 70t SPC. See statistical process control (SPC) spider charts, 253f constructing, 79–80 defined, 78–79 example of, 79f, 80–81 template, 277 sporadic variation, 183 stakeholder analysis, 10 approach for, 11 examples of, 14, 15–17 stakeholders defined, 10–11 expectations for improvement of, 9–17 identifying, 11 Kano model of, 12–14, 13f model of, 12f models for classifying, 11–12, 12f organizational expectations of, 12–13 requirements of, 14 strategies of, 14–15, 15f standard deviation, 183–184 statistical process control (SPC). See also control charts basic statistics for, 183–186 example of, 202–203 process capability and, 200–202 variation and, 183 statistics flowcharts with, 58–62, 61f principles of, 183–186 variation, 183 strategic benchmarking, 225 strategies. See also strategy reviews, organizational emergent, 18 organizational, reviewing, 17–18 strategy maps, 22–24 example of, 24–25 strategy reviews, organizational, 17–18 tools for competitive force analysis, 20–22 strategy maps, 22–25 SWOT analysis, 18–20 stratified sampling, 114 streamlining, 167–176 eliminating bureaucracy, 168–169 eliminating redundancy, 169

296  Index example of, 175–176 reducing process cycle times, 172–174 value-added analysis, 169–172 Structured Analysis and Design Technique (SADT), 49 structured brainstorming, 158 support processes, 35, 35f surveys, 115–117 example of, 116–117 procedure for, 116 swim lanes, 51 SWOT (strengths, weaknesses, opportunities, and threats) analysis, 18–19 example of, 19–20 template, 276 systematic reengineering, 209 systematic sampling, 114

T tampering, 185 tools for analyzing performance shortcoming bottleneck analysis, 151–155 cause-and-effect charts, 128–131 critical incident, 123–125 five whys analysis, 132–133 histograms, 136–142 is–is not analysis, 148–151 matrix diagrams, 145–148 Pareto charts, 125–127 relations diagrams, 142–145 scatter charts, 134–136 for business process improvement model, 111t for collecting data about performance shortcoming check sheets, 117–119 problem concentration diagrams, 119–122 sampling, 113–115 surveys, 115–117 for generating/selecting ideas brainstorming, 157–159 brainwriting, 159–160 Crawford slip method, 159–161 nominal group technique, 163–165 paired comparisons, 165–166 six thinking hats, 161–163 for implementing improvements ADT analysis, 240–241 force field analysis, 245–249 tree diagrams, 241–243

for improvements benchmarking, 221–236 business process reengineering, 207–221 idealizing, 177–178 quality function deployment, 179–182 Six Sigma, 204–207 statistical process control/control charts, 182–203 streamlining, 167–176 need for, 107–108 TOPP (Productivity Program of the Technology Industry) program, 35 traditional flowcharts, 46–51 training, for improvement projects, 94 transformation phase, of BPR, 210, 215–216 tree diagrams, 241–243, 242f creating, 242 example of, 243 trend analysis defined, 76 example of, 77–78 procedures for, 76–77 template, 276 T-shaped matrix diagrams, 146, 146f

U u charts, for attribute data, 188 constructing, 195–196 unstructured brainstorming, 158

V value-added analysis, 169–172 variation chronic, 183, 185 sporadic, 183

W Welch, Jack, 29

X x and R charts, constructing, 191–193 X bar and R charts, constructing, 189–191, 191f X-shaped matrix diagrams, 146, 146f

Y Y-shaped matrix diagrams, 146, 146f